FOOD AND DRUG ADMINISTRATION



















                           ADVISORY COMMITTEE
















                           September 10, 2004


                               8:31 a.m.








                              Holiday Inn

                          Versailles Ballrooms

                           Bethesda, Maryland



                        P A R T I C I P A N T S


      Dornette Spell-LeSane, M.H.A., NP-C, Executive



      Jeffrey S. Borer, M.D., Acting Chairman

      Steve Nissen, M.D.

      Alan T. Hirsch, M.D.

      Thomas Fleming, Ph.D.

      Maria H. Sjogren, M.D.

      Jonathan Sackner-Bernstein, M.D.

      John R. Teerlink M.D.


      Susanna L. Cunningham, Ph.D.

      William R. Hiatt, M.D.

      Beverly H. Lorell, M.D.

      Thomas Pickering, M.D.

      Ronald Portman, M.D.


      Paul Watkins, M.D., Ph.D.

      Jose Vega, M.D.


      FDA Participants


      Dr. Mark Avigan

      Dr. Florence Houn

      Dr. Joyce Korvick

      Dr. Kathy Robie-Suh



                            C O N T E N T S


      AGENDA ITEM                                             PAGE


      Call to Order and Introductions - Jeffrey S. Borer,

      M.D., Acting Chair                                         5


      Conflict of Interest Statement - Dornette

      Spell-LeSane, NP-C, Executive Secretary                    8


      Welcome and Comments - Norman Stockbridge, M.D.,

      Acting Director, Division of Cardiovascular and

      Renal Drug Products, FDA                                  13


      Sponsor Presentation


      Introduction - Hamish Cameron, M.D., Vice

      President, Exanta                                         15


      Clinical Pharmacology - Troy Sarich, Ph.D.,

      Director, Clinical Pharmacology                           24


      Efficacy - Jay Horrow, M.D., Senior Director,

      Clinical Development                                      43


      Safety - Sunita Sheth, M.D., Senior Director,

      Clinical Development                                      97


      Benefit and Risk Anticoagulation - Jonathan L.

      Halperin, M.D., Mount Sinai Medical Center, New

      York                                                     163


      FDA Presentation


      Risk/Benefit Assessment - Ruyi He, M.D., Medical

      Officer, Division of Gastrointestinal and

      Coagulation Drug Products                                172


      Risk Management of Hepatotoxic Drugs - Kate

      Gelperin, M.D., M.P.H., Medical Epidemiologist,

      Division of Drug Risk Evaluation                         282


      Drug-Induced Liver Toxicity - Paul Watkins, M.D.,

      Verne S. Caviness Distinguished Professor of

      Medicine, Director, General Clinical Research

      Center, University of North Carolina

      Medical Center                                           301



                      C O N T E N T S (Continued)


      AGENDA ITEM                                             PAGE


      Questions from the Committee                              --


      Open Public Hearing                                      184


      Charge to the Committee - Joyce Korvick, M.D.,

      M.P.H., Acting Division Director, Division of

      Gastrointestinal and Coagulation Drug Products, FDA      318


      Committee Discussion                                     319


      Break                                                     --


      Committee Questions/Summary                              320


      Adjournment                                              414




                         P R O C E E D I N G S


                DR. BORER:  It's 8:30 and I'm going to


      call the meeting to order.  This is the Cardiovascular and


      Renal Drugs Advisory Committee


      meeting, and we will discuss New Drug Application


      (NDA) 21-686, proposed trade name Exanta


      (ximelagatran) by AstraZeneca, for the proposed


      indication of the prevention of venous thromboembolism in


      patients undergoing knee replacement


      surgery, the prevention of stroke and other


      thromboembolic complications associated with atrial


      fibrillation, and the long-term secondary


      prevention of venous thromboembolic events after


      standard treatment following an episode of acute


      venous thromboembolic event.


                We'll begin by introducing everybody at


      the table.  In this meeting, the Cardio/Renal


      Committee actually is advising the GI Division as


      well as Cardio/Renal, in fact, primarily the GI


      Division, so we have more people at the table than


      we sometimes do.  Maybe we can each say our name


      and what we're doing here, and we'll start with Dr.




      Vega on the far side.


                DR. VEGA:  I'm Jose Vega.  I'm the


      industry representative on the committee, and I'm


      from Amgen.


                DR. PICKERING:  Tom Pickering from


      Columbia Presbyterian Hospital in New York.


                DR. PORTMAN:  Ron Portman from the


      University of Texas in Houston.


                DR. HIATT:  Bill Hiatt, University of




                DR. LORELL:  Bev Lorell, Harvard Medical


      School, and also Guidant Corporation.


                DR. SACKNER-BERNSTEIN:  Jonathan


      Sackner-Bernstein, North Shore University Hospital


      in New York.


                DR. CUNNINGHAM:  Susanna Cunningham.  I am


      the consumer representative on the committee, and


      I'm from the University of Washington.


                DR. NISSEN:  I'm Steve Nissen.  I'm a


      cardiologist at the Cleveland Clinic.


                DR. WATKINS:  Paul Watkins.  I'm a


      hepatology consultant from University of North




      Carolina-Chapel Hill.


                DR. BORER:  Jeff Borer, cardiologist,


      Weill Medical College of Cornell University.


                MS. SPELL-LeSANE:  Dornette Spell-LeSane,


      Executive Secretary for the committee.


                DR. TEERLINK:  John Teerlink, University


      of California-San Francisco, and San Francisco VA


      Medical Center.


                DR. FLEMING:  Tom Fleming, University of




                DR. HIRSCH:  Alan Hirsch, cardiologist and


      vascular medicine specialist at the University of


      Minnesota and Minneapolis Heart Institute.


                DR. AVIGAN:  Mark Avigan, Office of Drug


      Safety at the FDA.


                DR. STOCKBRIDGE: I'm Norman Stockbridge,


      the Acting Director of the Division of Cardio/Renal


      Drug Products at FDA.


                DR. HOUN:  I'm Florence Houn.  I'm the


      Office Director for Drug Evaluation III.


                DR. KORVICK:  Joyce Korvick, Acting


      Director, Division of Gastrointestinal Coagulation




      Drug Products.


                DR. ROBIE-SUH:  Kathy Robie-Suh, Acting


      Deputy Director, Division of Gastrointestinal and


      Coagulation Drug Products.


                DR. BORER:  Thank you very much.


                We have many people at the table.  I'm


      going to remind everyone that when you speak, you


      should press the button on your microphone, and


      when you're done, turn it off, please, unless you


      want to say something because that's the only way


      I'm going to know that you want to if you press the


      button and I see the light.


                We'll go on to the conflict of interest


      statement.  Dornette Spell-LeSane, the Executive


      Secretary of the Cardio/Renal Drug Advisory


      Committee, will present the conflict of interest.


                MS. SPELL-LeSANE:  Good morning.  The


      following announcement addresses the issue of


      conflict of interest and is made part of the record


      to preclude even the appearance of such at this


      meeting.  Based on the submitted agenda and all


      financial interests reported by the committee




      participants, it has been determined that all


      interests in firms regulated by the Center for Drug


      Evaluation and Research present no potential for an


      appearance of a conflict of interest at this


      meeting, with the following exceptions:


                In accordance with 18 U.S.C. Section


      208(b)(3), full waivers have been granted to the


      following participants.  Please note that all of


      the consulting and speaking activities waived are


      unrelated to Exanta and its competing products:


                Dr. William Hiatt for consulting for two


      competitors for which he receives less than $10,001


      per year per firm;


                Dr. Thomas Pickering for serving on a


      competitor's advisory board for which he receives


      less than $10,001 per year;


                Dr. Ronald Portman for consulting for a


      competitor for which he receives less than $10,001


      per year;


                Dr. Thomas Fleming for consulting for four


      competitors, he receives less than $10,001 per year


      per firm;




                Dr. Sackner-Bernstein for consulting for


      the sponsor and a competing firm, he receives less


      than $10,001 per year per firm.  Also, for his


      Speaker Bureau activities for a competitor, he


      receives less than $10,001 to $50,000 per year;


                Dr. Jeffrey Borer for serving on a


      steering committee for a competitor, he receives


      less than $10,001 per year;


                Dr. Alan Hirsch for lecturing for the


      sponsor, for which he receives less than $5,001 per


      year.  For lecturing for three competing firms, he


      receives less than $5,001 per year for serving on


      two Speaker Bureaus, and from $5,001 to $10,001 for


      one Speakers Bureau.  Two consulting agreements for


      two competing firms, he receives less than $10,001


      per year for one consulting, and from $10,001 to


      $50,000 per year for the other.


                In accordance with 18 U.S.C. 208(b)(3), a


      limited waiver has been granted to Dr. Paul Watkins


      for serving on two advisory boards for a competing


      firm.  He receives less than $10,001 per year for


      one and greater than $50,000 per year for the




      other.  Under the terms of this limited waiver, Dr.


      Watkins will be permitted to participate in the


      committee's discussion of Exanta.  He is, however,


      excluded from voting.


                Lastly, in accordance with 18 U.S.C.


      Section 208(b)(1), full waivers have been granted


      to the following participants for interests


      unrelated to Exanta and its competing products:


                Dr. John Teerlink for speaking for two


      competitors, he receives less than $10,001 per year


      from one, and from $10,001 to $50,000 per year from


      the other.  Also, for his consulting for a


      competitor for which he receives between $10,000 to


      $50,000 per year;


                Dr. Maria Sjogren for consulting for a


      competitor for which she receives less than $10,001


      per year.


                A copy of the waiver statement may be


      obtained by submitting a written request to the


      agency's Freedom of Information Office, Room 12A-30


      of the Parklawn Building.  In the event that the


      discussions involve any other products or firms not




      already on the agenda for which FDA participants


      have a financial interest, the participants are


      aware of the need to exclude themselves from such


      involvement, and their exclusion will be noted for


      the record.


                We would also like to note that Dr. Jose


      Vega has been invited to participate as an industry


      representative acting on behalf of regulated


      industry.  Dr. Vega is employed by Amgen.


                With respect to all other participants, we


      ask in the interest of fairness that they address


      any current or previous financial involvement with


      any firm whose products they may wish to comment




                Thank you.


                DR. BORER:  Thank you very much, Dornette.


      That was about the longest conflict of interest


      statement that I can remember.


                But we still are five minutes ahead,


      Norman, so we'll hear a welcome and comments from


      Norman Stockbridge, the Acting Director of the


      Division of Cardiovascular and Renal Drug Products.




                DR. STOCKBRIDGE:  I'll see if I can keep


      us on schedule.  Good morning and welcome to what


      promises to be an interesting meeting on behalf of


      the Divisions of Cardio/Renal Drug Products and GI


      and Coagulation Drug Products.  I want to thank


      members of the Cardio/Renal Advisory Committee,


      consultants, and the sponsor for their




                I do need to acknowledge retirement of


      four members from the Advisory Committee:  Alan


      Hirsch is here today, a couple of chairs down to my


      left; Steve Nissen is over at the middle of the


      table there; Paul Armstrong would be here today


      except that Homeland Security discovered that he's


      a Canadian.




                DR. STOCKBRIDGE:  And, finally, there is


      our Chairman, Dr. Jeff Borer.  Dr. Borer's service


      to the committee began in 1977, an era in which


      members still sported powdered wigs.




                DR. STOCKBRIDGE:  I can't quite tell from




      the records where he cast his first vote, but in


      that year, the committee heard arguments on


      potassium and atropine.


                As tokens of our appreciation, Ms.


      Spell-LeSane has for each of you some actual


      certificates signed by our Acting Commissioner and


      some virtual plaques that look just like this one.


      So on behalf of Cardio/Renal, the Food and Drug


      Administration, and a grateful nation, thanks to


      you all.




                DR. NISSEN:  Norman, I'm not from Canada,


      but I'm from Cleveland, and it's really close to


      Canada.  Will you please not tell Homeland Security


      about me?


                DR. BORER:  Thank you very much, Norman,


      and thank you for staying way on time because we


      are now 17 minutes ahead of schedule, which is


      good.  The sponsor has a 90-minute presentation.


      We'll try to allow you to move along as well as we


      can, but undoubtedly there will be some clarification


      questions.  We ought to try to hold the




      questions that we ask to clarification issues


      during the presentation, if we can, and we can get


      into the meat of the substantive discussion




                The presentation will be introduced by Dr.


      Cameron, the Vice President of Exanta.


                DR. CAMERON:  Thank you, Mr. Chairman,


      members of the committee, ladies and gentlemen,


      good morning.  I'm Dr. Hamish Cameron, the Vice


      President of Exanta at AstraZeneca, and with my


      colleagues we're pleased to present ximelagatran, a


      new oral anticoagulant.


                After a 20-year journey to discover and


      develop this new medicine and half a century


      without significant innovation in this area of


      therapeutics, we believe ximelagatran, the first


      oral treatment in the new drug class direct


      thrombin inhibitors is a real advance in oral




                Ximelagatran has a mechanism of action


      that's quite different from the vitamin K


      antagonists like warfarin and can provide the first




      oral alternative to warfarin, today's only option


      for long-term anticoagulation.


                Anticoagulation is the major approach to


      both the prevention and treatment of thromboembolic


      disease, a disease that's the final common pathway


      for many life-threatening conditions, like stroke,


      myocardial infarction, and pulmonary embolism.  And


      it's the commonest cause of death and disability in


      America today.


                At the outset we must ask:  Given the


      widespread availability of the vitamin K


      antagonists like warfarin, why is there a need for


      a new oral anticoagulant?  Warfarin is a highly


      efficacious anticoagulant and one of the top ten


      most prescribed drugs, used in nearly every medical


      specialty by 3 million patients in the U.S.


      involving 32 million prescriptions every year.  But


      it's been in the top five, sometimes number one, in


      the lists of drugs associated with significant


      interactions, medication errors, serious bleeding,


      and hospital admissions.


                Warfarin's profile of unpredictable




      kinetics and dynamics; food, alcohol, and multiple


      drug interactions; together with its acknowledged


      narrow therapeutic index--too little warfarin, and


      there's the risk of residual clotting; too much,


      and the risk of bleeding--all these drive the need


      for a lifetime of INR coagulation monitoring and


      never-ending individual dose titration.


                To put it simply, you don't get the


      benefit of warfarin from just taking the tablet.


      Its overall effectiveness is highly dependent on


      how it's managed.  And it's this fact that frames


      the innovation of ximelagatran.


                Many patients and doctors fear the risk of


      bleeding that comes with unpredictable anticoagulation.


      This fear tends to result in


      under-treatment, quite paradoxical in high-risk


      elderly patients, or in about half the overall


      patients eligible for warfarin, little or no


      treatment at all.


                We started a discovery program targeting


      thrombin in 1985.  We sought to develop a new oral


      anticoagulant, an alternative to warfarin, with a




      profile that would allow fixed dosing without


      coagulation monitoring, further supported by a low


      potential for food and drug interactions.


                We looked for a rapid onset and offset of


      action to simplify turning anticoagulation on and


      off, which is one of the challenging aspects of


      warfarin treatment.  And all this had to be


      achieved with an acceptable bleeding profile.


                Today, we believe these objectives have


      been met by the Ximelagatran Development Program,


      involving 82 clinical studies and enrolling over


      30,000 subjects.  More than 17,000 people received


      ximelagatran with 3,500 patients dosed for over a


      year.  And our longest patient exposure has now


      reached five years.


                Here are the three proposed indications in


      the current NDA, spanning exposures from days to


      several years.  The first is the long-term


      secondary prevention of venous thromboembolism,


      VTE, after standard treatment for an acute episode.


      Treatment of acute VTE involved six months of


      anticoagulation with warfarin, but at the start of




      this development in 1999, it was unknown whether


      longer treatment would be beneficial.  And so a


      placebo-controlled study, THRIVE III, was


      conducted.  As you'll see, this study demonstrated


      a highly significant reduction of VTE during


      longer-term prophylactic treatment.  And as a


      placebo-controlled study, it provides the strongest


      evidence of ximelagatran's antithrombotic efficacy.


                The second indication is the prevention of


      VTE after knee replacement surgery.  Patients


      without anticoagulant prophylaxis run a high risk


      of DVT and pulmonary embolism, and in the U.S.,


      warfarin is the most widely used drug, started late


      on the day of surgery to reduce this risk.  In two


      warfarin controlled studies, EXULT A and EXULT B,


      we've shown a significant reduction of VTE risk for


      ximelagatran compared with warfarin.


                The third indication is the prevention of


      stroke and other thromboembolic complications


      associated with atrial fibrillation.  Here with the


      large SPORTIF III and V trials, we've shown the


      efficacy of ximelagatran to be comparable to




      well-controlled warfarin.


                Across these pivotal, mainly


      outcomes-based studies involving independent


      endpoint adjudication, we've demonstrated


      ximelagatran's antithrombotic efficacy and recorded


      a favorable leading profile, equivalent to and in


      some cases better than the comparator.  We detected


      a signal of raised hepatic enzymes with chronic


      treatment, and so we've conducted a very detailed


      analysis, consulted with experts, and believe the


      risk can be adequately managed.


                I should highlight that in your briefing


      packs and our safety presentation, we've included


      data from two other large studies in other


      indications, not for consideration today, but which


      contribute nearly 4,000 patients.  The THRIVE


      treatment study is the first pivotal study looking


      at initial VTE treatment, and the second study is


      soon to start, while the ESTEEM trial is a Phase II


      dose guiding study in the post-acute coronary


      syndrome setting.


                These data enrich the overall safety




      assessment, including patients from very different


      clinical settings and a wide range of characteristics, and


      we're going to review all the key


      data in the presentations that follow.


                We believe ximelagatran with its


      predictable anticoagulant effects and favorable


      bleeding profile has a positive benefit/risk in the


      proposed indications, provided it's used properly.


      And part of that proper use is the introduction of


      an appropriate risk management program directed


      towards the hepatic risk.


                We made an initial proposal with our


      submission which had been developed with extensive


      external consultation and field testing, but we


      fully recognize, following the deliberations of


      this committee and further discussions with FDA,


      the program will need to be developed and


      strengthened further before an approach can be


      finalized in the best interests of patients.  We


      are committed to working with FDA to achieve the


      most appropriate risk management program to ensure


      the safe use of ximelagatran because patient safety




      is, has been, and always will be AstraZeneca's top


      priority when we introduce new medicines into


      clinical practice.


                Since 1998, we've met repeatedly with FDA


      throughout ximelagatran's development--in end of


      Phase II meetings, a pre-NDA interaction, and


      there's a meeting coming up to discuss the nature


      and extent of the risk management program.  The NDA


      was submitted in December 2003.


                I should add that all the same data are


      now being reviewed in Europe by the French agency


      before a mutual recognition procedure.  But there's


      one difference worth noting.  Given the quite


      different clinical practice regarding anticoagulant


      prophylaxis in orthopedic surgery, separate


      developments were conducted in the U.S. and Europe.


      The European program, reflecting local practice and


      starting treatment much closer to the time of


      operation, was the subject of a separate earlier


      regulatory submission and completed the mutual


      recognition procedure in May this year.  And the


      first orthopedic launch was in Germany in June.




                Now, you have the data on this program in


      your briefing packs, but with the significant


      timing, comparator, and formulation differences,


      our presentations today will largely focus on the


      data directly relevant to the NDA orthopedic


      surgery application.


                Here's the agenda for our session.  Dr.


      Troy Sarich will review clinical pharmacology; Dr.


      Jay Horrow, efficacy; and Dr. Sunita Sheth, safety;


      allowing an overall evaluation of ximelagatran in


      the three requested indications.  Dr. Jonathan


      Halperin from Mount Sinai Medical Center will then


      give his views on the benefit/risk of ximelagatran


      in clinical practice.  And throughout our


      presentations, we hope to cover for you all the


      specific comments raised by the agency in their


      briefing document.


                In addition to Dr. Halperin, we're also


      joined by other consultants:  Dr. Gerald Faich, Dr.


      Lloyd Fisher, Dr. Peter Kowey, and Dr. James Lewis.


                In summary, then, ximelagatran is a new


      oral anticoagulant that provides the first




      alternative to warfarin after 50 years.  We believe


      a total review of the available clinical data


      supports a positive benefit/risk in each of the


      proposed indications.  Ximelagatran can enhance


      health care delivery in America and throughout the


      world to help prevent a range of debilitating and


      life-threatening thromboembolic diseases.


                Thank you.  Now I'd like to introduce Dr.


      Troy Sarich for clinical pharmacology.


                DR. SARICH:  Good morning.  I'm Dr. Troy


      Sarich, Director of Clinical Pharmacology at


      AstraZeneca.  I'll now present an overview of the


      clinical pharmacology of ximelagatran in which we


      have performed both the traditional clinical


      pharmacology studies and population pharmacokinetic




                Ximelagatran is an oral direct thrombin


      inhibitor.  It's rapidly bioconverted to the active


      form, melagatran.  The bioconversion, which is


      Cytochrome P-450 independent, involves both


      de-esterification and a reduction that occurs


      throughout the body.




                The exposure to melagatran is linear


      across a dose range from 5 to 98 milligrams


      ximelagatran.  The pharmacokinetics are predictable


      over time with repeated dosing, and the elimination


      half-life of melagatran is approximately 4 to 5


      hours in patients.  Once formed, melagatran is


      primarily eliminated from plasma by a glomerular




                Thrombin is a key enzyme in the


      coagulation cascade.  It converts fibrinogen to


      fibrin, activates platelets, and induces its own


      generation.  Melagatran directly inhibits thrombin


      as a classic competitive and reversible-binding


      enzyme inhibitor.  There's a direct relationship


      between the pharmacokinetics and pharmacodynamics


      of ximelagatran.  Its active when present in


      plasma, and once eliminated from plasma, its effect


      is gone.


                Preclinical investigations indicated an


      antithrombotic effect of melagatran at approximately 0.05


      micromolar, with increasing effect up


      to approximately 0.5 micromolar.




                In humans, ximelagatran prolongs clotting


      time assays.  The thrombin time assay shown here


      was prolonged in a linear manner at concentrations


      as low as 0.05 micromolar.  In addition, melagatran


      prolongs in a concentration-dependent manner the


      activated partial thromboplastin time, although it


      is less sensitive.


                Additional investigations using


      pharmacodynamic models in humans demonstrated


      evidence for inhibition of thrombin generation


      indicated by concentration-dependent reduction and


      thrombin-antithrombin complex levels and platelet


      activation indicated by concentration-dependent


      reduction in beta thromboglobulin levels at


      melagatran concentrations at or near 0.05


      micromolar.  All together, it was clear from these


      data that direct inhibition of thrombin by


      melagatran resulted in the intended anticoagulant


      activity in humans.


                After oral administration, the inactive


      pro drug ximelagatran is rapidly eliminated from


      plasma as it is biotransformed to melagatran, shown




      in blue, with peak melagatran concentrations


      occurring approximately 2 to 3 hours post-dosing.


      Melagatran plasma concentrations greater than 0.05


      micromolar are achieved early after oral ximelagatran,


      indicating a rapid onset of action which


      simplifies the initiation of oral anticoagulation.


      Concentrations remain above 0.05 micromolar


      throughout the dosing interval, supporting a


      twice-daily dosing regimen.


                And as shown here, the rapid onset of


      action of oral ximelagatran is not altered when


      co-administered with food.  Although there's an


      approximately one-hour delay in the time to C-max,


      there is no effect on the AUC or Cmax of




                Warfarin's well-recognized drug


      interaction profile is largely related to its


      metabolism by the Cytochrome P-450 system and its


      high plasma protein binding.  Ximelagatran is not


      metabolized by and does not inhibit the major


      Cytochrome P-450 enzymes listed here.  It also has


      low plasma protein binding, and along with the




      majority of melagatran eliminated from plasma by


      glomerular filtration, this leads to an inherently


      low potential for drug interactions.


                Our investigations have identified


      pharmacokinetic interactions with erythromycin and


      azithromycin.  Erythromycin results in an


      80-percent or less than twofold increase in


      melagatran plasma levels, with a smaller, 40- to


      60-percent increase with azithromycin.


                These changes are within the overall range


      of melagatran exposures in patients, and as


      outlined in detail in your briefing document,


      investigation into the potential impact of this


      pharmacokinetic interaction found no signal for


      increases bleeding events or increased ALT


      elevations in the approximately 230 patients


      receiving ximelagatran and macrolide antibiotics in


      the long-term studies.  These data do not suggest


      an important clinical impact of these


      pharmacokinetic interactions.


                We have conducted many other interaction


      studies where we've found no significant




      pharmacokinetic interactions.  As shown by the mean


      melagatran AUC ration and the 90-percent confidence


      interval within or slightly outside the 0.8 to 1.25


      no interaction interval.  The drugs investigated


      include alcohol, common cardiovascular medications,


      an NSAID, a sedative, and several antibiotics.


      These results are consistent with population


      pharmacokinetic analyses indicating a lack of


      interaction with commonly used comedications in the


      patient studies.  Taken together, these data


      suggest that ximelagatran has a low potential for


      drug interactions.


                Melagatran is primarily renally eliminated


      from plasma, and so we've carefully investigated


      the impact of renal function on the pharmacokinetics of


      ximelagatran.  In the three patient


      populations under consideration today, melagatran


      exposure increases as calculated creatinine


      clearance decreases.  For this reason, severe renal


      impairment, a calculated creatinine clearance less


      than 30 mLs per minute, was an exclusion criteria


      for our clinical studies, and we're currently




      investigating an alternative dosing strategy in


      that population.


                It's notable that we've gathered


      considerable experience with ximelagatran in


      patients with mild to moderate renal impairment as


      approximately 45 percent of the Phase III patient


      population had a calculated creatinine clearance


      between 30 and 80 mLs per minute.  The median


      exposures in these patients are about 1.5 to 2.5


      times higher, respectively, than patients with


      normal renal function, but there's considerable


      overlap in melagatran exposure between groups,


      suggesting dose adjustment was not necessary.


                We've also studied the potential effects


      on the pharmacokinetics of ximelagatran within


      other special populations, and other than


      differences in renal function between groups, we


      have not identified other important effects of age,


      gender, race, obesity--as measured using body mass


      index--or body weight on the pharmacokinetics of




                The agency has suggested there should be a




      dose adjustment for ximelagatran in patients with


      renal impairment given the higher levels of


      melagatran in these patients.  But I'd like to show


      you why a fixed dose, as used in our clinical


      studies, is appropriate across the patient


      populations studied.


                We do agree with the agency's assessment


      that there's no need for dose adjustment in


      orthopedic surgery patients and that there was no


      increased bleeding related to melagatran exposure


      in VTE secondary prevention patients.  We do


      acknowledge an association between increasing


      melagatran exposure and increasing incidence of


      major bleeding in atrial fibrillation patients.


      But this relationship appears confounded by the


      correlation between melagatran exposure and the


      age-related decrease in calculated creatinine




                Shown here from the SPORTIF trials is the


      relationship between calculated creatinine


      clearance and major bleeding.  As you can see,


      major bleeding increased with declining renal




      function whether patients received ximelagatran or


      INR-controlled warfarin.  This suggests the


      increase in melagatran concentrations in patients


      with renal impairment is not associated with


      increased bleeding versus INR-controlled warfarin.


                It's also important to note that stroke


      risk increased with decreasing calculated


      creatinine clearance, and the vast majority of


      these strokes were ischemic.  So there's a


      possibility that a dose reduction in renally


      impaired patients intended to decrease bleeding may


      increase the risk of stroke in those patients at


      highest risk.


                We should also consider the hepatic


      findings, as will be presented by Dr. Sheth, and we


      have examined the possible relationship between


      melagatran exposure and ALT elevations.  As pointed


      out in our briefing document, we have observed an


      association between increasing melagatran exposure


      and increasing ALT elevations greater than 3 times


      upper limit of normal, but this relationship is


      very weak.  And as shown here, the relationship




      between melagatran AUC and peak ALT elevation in


      individual patients, while statistically


      significant, does not suggest a clear relationship


      between melagatran exposure and ALT elevations.


                In addition, we agree with the agency's


      conclusion that, aside from the ALT elevations


      noted with ximelagatran, there is no difference in


      the overall adverse event profile between


      ximelagatran and comparators in the long-term


      dosing study pool.


                Factoring in the occurrence of major


      bleeding, stroke and systemic embolic events, ALT


      elevations, and the overall adverse event profile,


      the observation of increased plasma melagatran


      concentrations in renal impairment does not appear


      to justify a dose reduction in these patients.  We


      believe our data support a fixed dose of


      ximelagatran in the patient populations studied.


                Now I'd like to show the steady-state


      plasma concentrations of melagatran in atrial


      fibrillation patients receiving a fixed dose of 36


      milligrams ximelagatran twice daily.  There are




      four key points here.


                Plasma concentrations fluctuate during the


      dosing interval, remaining largely above 0.05


      micromolar and infrequently exceeding 1 micromolar.


                Mean trough melagatran concentrations


      after 36 milligrams are approximately 0.2


      micromolar.  So should a patient miss a dose of


      ximelagatran, the 4- to 5-hour half-life of


      melagatran means that low but pharmacologically


      active concentrations remain for up to 24 hours


      post-dosing.  And the effect of melagatran is gone


      once it is cleared from plasma by the kidneys.


                This emphasizes the importance of


      maintaining good diuresis in the management of


      bleeding.  And while there is no specific antidote,


      if needed, melagatran can be dialyzed.


                And, lastly, the APTT is prolonged in


      patients and may help identify a residual


      anticoagulant effect.


                A critical aspect of oral anticoagulation


      is maintenance of a stable effect over time, and we


      have confirmed the long-term stability of oral




      ximelagatran.  Shown in yellow are the plasma


      concentrations of melagatran in atrial fibrillation


      patients in the Phase II study, SPORTIF II.  We


      remeasured plasma melagatran concentrations in a


      subset of those same patients between 13 to 16


      months later in SPORTIF IV, a long-term


      continuation study of SPORTIF II.


                The mean plasma concentrations of


      melagatran are completely overlapping, and the


      variability in exposure within individual patients


      was low, with a coefficient of variation of 25


      percent, indicating that oral ximelagatran results


      in stable and reproducible plasma concentrations of


      melagatran with long-term repeated dosing.  This


      stability enabled us to conduct our clinical


      studies using a fixed dose without coagulation




                So we can conclude from this extensive


      clinical pharmacology program pharmacologically


      active concentrations of melagatran are rapidly


      achieved and maintained in a broad range of


      patients.  There is also no effect of food or




      alcohol and a low potential for drug interactions.


                The key attributes of ximelagatran are,


      therefore, its oral availability, rapid onset of


      action, low potential for drug interactions, and


      use at a fixed dose without coagulation monitoring.


                Now I'd like to introduce Dr. Jay Horrow,


      who will provide to you an evaluation of the


      efficacy of ximelagatran demonstrated in Phase III


      clinical studies.


                DR. BORER:  We'll just stop for one moment


      to make sure there are no issues that need to be


      clarified.  The relation of renal function to


      melagatran exposure undoubtedly is going to be


      discussed to a greater extent later, but I think we


      should hold that until we hear from the FDA


      presentations, and then we can talk about that.


      But if there are any issues that need to be


      clarified regarding the pharmacology, we should do


      that now.




                DR. NISSEN:  Two very brief questions.  Is


      anything known about the mechanism of macrolide




      interaction?  Have you explored that at all?


                DR. SARICH:  Yes.  We were slightly


      surprised to find that interaction since we don't


      interact with the P-450 system.  It appears that


      the interaction involves transport proteins of some


      kind, and we've looked at a range of different


      compounds that we've investigated, and it at this


      point appears isolated to the macrolide antibiotics


      we've studied.


                DR. NISSEN:  And the second question is:


      You showed the coagulation effect during therapy,


      and I wondered if you have additional data on what


      happens in, let's say, the first 72 to 96 hours


      after terminating therapy.  Is there evidence of a


      rebound phenomenon?


                DR. SARICH:  We have not observed that


      pharmacologically, as far as coagulation time




                DR. NISSEN:  Okay.  But that has been




                DR. SARICH:  We've followed out to 24


      hours after single-dose administration and not seen




      any evidence--


                DR. NISSEN:  But not longer than 24 hours?


                DR. SARICH:  Not that I can recall.


                DR. BORER:  John?


                DR. TEERLINK:  The other question I have


      is:  In terms of the relationship between the


      melagatran AUC versus the peak ALT elevations, how


      was the melagatran AUC derived?


                DR. SARICH:  Yes, these were derived using


      a population pharmacokinetic model.  So the


      patients that received ximelagatran in the Phase


      III clinical studies had plasma samples collected.


      Over 80 percent of the Phase III patient


      population--in the long-term population had a


      plasma sample collected.  Using a pharmacokinetic


      model that was developed by the team, we were able


      to estimate the exposure to melagatran in those




                DR. BORER:  Ron Portman?


                DR. PORTMAN:  Noting differences in the


      chronopharmacology of drugs, were the curves you


      showed similar for both the morning and evening






                DR. SARICH:  Are you speaking about the


      coagulation time assay--


                DR. PORTMAN:  No.  I was talking about the


      plasma concentrations.


                DR. SARICH:  Pharmacokinetics?


                DR. PORTMAN:  Right, pharmacokinetics.


                DR. SARICH:  Yes, they are consistent


      under administration during the day or overnight.


                DR. BORER:  Jonathan, go ahead.


                DR. SACKNER-BERNSTEIN:  In the analysis


      that you showed the stability of the concentrations


      of the drug over time from the SPORTIF II and


      SPORTIF IV population, did you perform that


      analysis restricting to patients who had samples at


      both times?  Because the analysis you showed had a


      larger population at baseline compared to a subset




                DR. SARICH:  Right.  We've done it both


      ways.  The figure actually represents the larger


      number in the SPORTIF II study and a smaller number


      in SPORTIF IV.  The intra-subject variability I




      noted was only the subjects that had sampling at


      both time occasions.


                DR. BORER:  Alan?


                DR. HIRSCH:  In the PK and AUC curves that


      you've generated, were there any changes or


      differences noted based on ethnicity, geographic


      sampling of a population, or gender?


                DR. SARICH:  Are you asking pharmacokinetic--


                DR. HIRSCH:  Yes, PPK differences between




                DR. SARICH:  The main factor we've


      observed between any subgroups has been differences


      in renal function, calculated creatinine clearance.


      We have not observed any significant effects of


      other demographic parameters, age, gender, race,


      BMI, body weight.  It appears that exposure--the


      most influential demographic factor is calculated


      creatinine clearance.


                DR. BORER:  Susanna?


                DR. CUNNINGHAM:  Did you have a sufficient


      African American population to actually know




      anything about what the African American area of


      the curve might be or handling of the drug?


                DR. SARICH:  We have performed pharmacokinetic


      studies in that population.  I should say


      both--I'll show you some data here from a small


      study.  It's not African Americans per se, but it


      was a study in Europe, in Paris, in fact, where we


      had 12 blacks, 12 Asians, and 12 Caucasians, and


      found no real differences between these groups.


                If we looked at the entire patient


      population, we can see here--if we look at--you can


      see the Caucasian population here.  There's over


      6,000 patients.  The blacks where we had


      appropriate pharmacokinetic information, were 115,


      as well as Asians, and the category of other, and


      no differences between these populations.


                DR. BORER:  Tom?


                DR. PICKERING:  Do you have any data on


      interaction with aspirin?


                DR. SARICH:  Yes, we have performed


      actually two studies with aspirin.  There's no


      pharmacokinetic interaction with aspirin.  We see




      an additive effect on the capillary bleeding time,


      which is somewhat expected.


                DR. BORER:  Beverly?


                DR. LORELL:  Yes, with regard to body


      size, you commented on and emphasized obesity.


      What about the other end of the scale, very small


      body size?  Sometimes an issue in elderly women who


      might be candidates for several of these




                DR. SARICH:  We have less data in very


      small individuals, but what we know about that


      population is that it's primarily their calculated


      creatinine clearance that influences their




                DR. BORER:  Jonathan?


                DR. SACKNER-BERNSTEIN:  I know we're going


      to get back to the renal function question, but


      there was one set of slides you showed where you


      tried to give us some reassurance about the


      relationship between bleeding and renal function.


      And you showed the risk of bleeding as calculated


      creatinine clearance reached the low end of the






                I wonder if you performed any sort of


      retrospective power calculation on your ability to


      detect a difference in risk, in particular in the


      patients who we may be likely to see treated with


      this drug in clinical practice, those over 70, over


      75, where calculated creatinine clearances often


      are in the 40s.  So do you have an analysis there


      between 30 and 50 with conditional power to


      actually detect a difference in bleeding risk




                DR. SARICH:  I think we could probably


      best address that after the presentations.  We do


      have data there, and rather than getting into that


      discussion, if the Chair would agree, we could


      address that, bring an answer to you for that.


                DR. BORER:  Is that okay, Jonathan?


                DR. SACKNER-BERNSTEIN:  Yes.


                DR. BORER:  Okay.  Thank you.


      T1B                      DR. HORROW:  Ladies and


      gentlemen, I'm Dr. Jay Horrow from AstraZeneca.  We


      will now present Phase III data demonstrating that




      ximelagatran is an effective oral anticoagulant.


                In the first indication, long-term


      secondary prevention of venous thromboembolism, we


      will show ximelagatran superior to placebo.  In the


      second indication, prevention of VTE after total


      knee replacement, ximelagatran was superior to


      well-controlled anticoagulation with warfarin.  And


      in the chronic prevention of stroke, ximelagatran


      was noninferior to warfarin.


                These indications represent a broad range


      of patient populations.  We'll begin with the first


      one:  secondary prevention of VTE.


                Evidence has been accumulating that


      patients with acute VTE benefit from prolonged


      anticoagulation after acute treatment.  The THRIVE


      III trial comparing ximelagatran to placebo


      contributes to this growing body of evidence.


      Randomized patients had an acute symptomatic VTE


      objectively confirmed and had completed 6 months of


      treatment without VTE recurrence, also objectively


      documented at randomization.  Anticoagulation was


      desirable but not essential for these patients,




      that is, they had idiopathic VTE or probable


      hypercoagulable conditions.  Health status had to


      be compatible with survival for an additional 18




                In THRIVE III, 1,223 patients receives in


      double-blind fashion either oral ximelagatran 24


      milligrams twice daily or placebo for up to 18


      months.  Selection of 24 milligrams for this trial


      came from a consideration of preclinical data and


      data from Phase II trials in the orthopedic surgery


      indication.  These PK data from a Phase II European


      trial in patients undergoing hip or knee


      replacement demonstrate that administration of 8


      milligrams ximelagatran twice daily, the lowest


      curve, achieves plasma melagatran concentrations of


      about 0.05 micromolar.  This is the level at which


      anticoagulant activity with melagatran begins based


      on the data previously shown by Dr. Sarich.


                Progressively higher doses of oral


      ximelagatran, 12, 18, and 24 milligrams, achieved


      higher melagatran concentrations, more anticoagulant


      activity, and more time above the 0.05




      micromolar threshold for each dose.


                Outcome data from that same orthopedic


      surgery trial suggest that 24 milligrams is the


      most promising dose for efficacy.  The 24-milligram


      dose also had a reassuring bleeding profile.  We


      chose 24 milligrams for THRIVE III with placebo


      comparator without establishing dose-limiting


      toxicity, in this case bleeding.  The choice was an


      informed judgment taking into consideration, first,


      the need for efficacy demonstrated by the benefits


      seen here in joint replacement, an intense


      thrombotic stimulus; and, second, the need to avoid


      excess bleeding because the standard of care is no


      anticoagulant therapy at all.


                The trial compared ximelagatran to placebo


      in the rate of recurrence of symptomatic,


      objectively confirmed VTE.  VTE encompasses both


      deep vein thrombosis, DVT, and pulmonary embolism,


      PE, because PE originates from a thrombus in the


      systemic venous circulation, whether overt or not.


                The primary endpoint compared ximelagatran


      to placebo using a time-to-event analysis.  A




      recurrence of VTE required signs or symptoms of


      VTE, that is, a clinical event, and subsequent


      objective confirmation.  A blinded independent


      endpoint committee evaluated and adjudicated all


      clinical endpoints, including major bleeding




                The ximelagatran- and placebo-treated


      cohorts displayed similar demographic profiles.  As


      indicated by creatinine clearance between 30 and


      80, 23 percent had some degree of renal impairment.


      The index VTE event was or included pulmonary


      embolism for more than one-third of patients.  This


      Kaplan-Meier curve shows the cumulative incidence


      of the primary outcome in the ximelagatran and


      placebo groups, analyzed by intention to treat.


      Seventy-one patients in the placebo group suffered


      recurrent VTE, including 23 PEs, for a cumulative


      rate of 12.6 percent, while only 12 patients in the


      ximelagatran group had recurrent VTE, including


      only two PEs, for a cumulative rate of 2.8 percent.


      The 9.8-percent difference, significant at p less


      than 0.0001 by log rang test indicates that one VTE




      recurrence is prevented by ximelagatran treatment


      for up to 18 months to 10 patients.  The associated


      hazard ratio, 0.16, indicates a risk reduction of


      84 percent by ximelagatran relative to placebo.


                The composite endpoint of total VTE


      included both DVT and PE.  Benefit of ximelagatran


      over placebo occurred for each component of this


      composite endpoint--clinical DVT, clinical PE, and


      their combination.


                The superiority of ximelagatran to placebo


      is robust to multiple, prespecified sensitivity


      analyses listed here.  Each comparison demonstrated


      a significance level less than 0.0001


                Here we examine efficacy in subpopulations.  Small


      diamonds depict point estimates of


      the odds ratios of ximelagatran to placebo, and


      horizontal bars show their 95-percent confidence


      intervals.  Superiority of ximelagatran over


      placebo remains in all subgroups strata of


      reasonable size.


                In THRIVE III, the oral thrombin inhibitor


      ximelagatran, 24 milligrams twice daily for up to




      18 months, effectively reduced the number of


      recurrent VTE events following 6 months' treatment


      of an acute VTE.  The results are robust and


      consistent across multiple endpoints and subgroups


      and demonstrate a clinically relevant benefit.


                The second indication under review today


      is the prevention of VTE in patients undergoing


      knee replacement surgery.  Major joint replacement


      surgery challenges any anticoagulant to prevent VTE


      without counteracting surgical hemostasis.  VTE


      prevention contributes heavily to the benefit/risk


      balance for joint replacement surgery.


                The current options to reduce the


      occurrence of VTE after total knee replacement


      include the injectable agents low-molecular-weight


      heparin and fondaparinux and oral warfarin.  One


      FDA comment regards the choice of warfarin as the


      comparator for these trials.


                We chose warfarin for several reasons:


                First, it is the agent most commonly used


      for this purpose in North America, and we


      administered warfarin, as orthopedic surgeons do,




      beginning the night of surgery.


                Second, warfarin, like


      low-molecular-weight heparin, is a Grade 1A


      recommended therapy for this purpose, according to


      current American College of Chest Physicians


      Consensus Conference Guidelines.


                And, third, warfarin is associated with


      less bleeding than the injectable anticoagulants


      and so is a more daunting comparator for


      ximelagatran in terms of surgical hemostasis.


                Two independent double-blind Phase III


      trials--EXULT A and EXULT B--enrolled patients


      undergoing primary elective total knee replacement.


      EXULT A studied 24 and 36 milligrams ximelagatran


      and warfarin.  We had studied 24 milligrams in this


      context previously and found protection similar to


      but not better than warfarin at p equal 0.07.  We


      found that result surprising.  Unsure whether or


      not it was a Type II error, we designed EXULT A


      with two ximelagatran arms:  one using 24


      milligrams and the other using 36 milligrams.


                Warfarin and its paired placebo began, as




      typically practiced in the U.S., the evening of the


      day of surgery while ximelagatran and its placebo


      began early on the morning after the day of


      surgery.  Treatment continued for 7 to 12 days,


      after which all patients underwent bilateral




                Based on the results of EXULT A, EXULT B


      studied only 36 milligrams ximelagatran and


      warfarin.  Warfarin was aggressively and


      successfully dosed to drive the INR rapidly to its


      target of 2.5, with an accepted range of 1.8 to


      3.0.  The primary outcome formed the composite of


      distal and proximal DVT by venogram performed


      between days 7 and 12, objectively confirmed


      symptomatic DVT or pulmonary embolism up to 2 days


      after venography, and all-cause mortality up to 2


      days after venography.  Both trials utilized the


      same blinded independent committee for event




                The treatment groups in the EXULT trials


      were balanced and represented well the population


      of patients in the United States undergoing total




      knee replacement.  More than a third of the cohort


      displayed some degree of renal impairment.


                Here are the primary results for EXULT A


      and EXULT B.  In EXULT A, ximelagatran 36


      milligrams, in yellow, showed superiority to


      well-controlled anticoagulation with warfarin, in


      gray, at p equals 0.003.  EXULT B confirmed those


      results, with p less than 10                                              

                               -5.  These results yield


      relative risk reductions of 26 and 29 percent and


      numbers needed to treat of 14 and 11, respectively.


                In EXULT A, ximelagatran 24 milligrams, in


      orange, and warfarin, in gray, did not differ, with


      event rates of 24.9 and 27.6 percent, respectively.


      That p value is 0.28.


                The delay in anticoagulation with warfarin


      administration suggests that it may act like a


      placebo in EXULT.  In fact, warfarin rates, in


      gray, are the lowest ever obtained in knee


      replacement clinical trials with warfarin, perhaps


      because of the rapid achievement in EXULT of


      therapeutic INRs.  Placebo rates are historically


      over 60 percent, and the mean INR in EXULT was 2.4




      on post-op day 3.  The warfarin group provided a


      formidable comparator for ximelagatran in the EXULT




                Here we see results for the components of


      the composite primary endpoint.  As expected, the


      majority of events occurred in the distal leg.


      Rates for proximal DVT, for PE, and for death were


      low in all treatment groups.  Another point raised


      by FDA is how clinically relevant distal DVT is as


      a component of that endpoint.  It's important to


      note that 10 to 20 percent of distal thrombi extend


      to become proximal thrombi, and either one can


      cause pulmonary embolism, making all three


      phenomena clinically relevant components of a


      composite endpoint.  In fact, proximal and distal


      deep vein thrombosis detected by venography,


      whether symptomatic or not, is a primary endpoint


      historically accepted by the agency for VTE


      prophylaxis registration trials.


                This display of the primary outcome by


      subgroup strata shows differences in event


      incidences between the pooled 36-milligram




      ximelagatran and pooled warfarin groups.  These


      results, with small numbers in just a few


      subgroups, reveal no discrepancies in efficacy in


      any particular subpopulation.


                Oral ximelagatran, 36 milligrams, provided


      superior protection against VTE and all-cause


      mortality compared with well-controlled


      anticoagulation with adjusted-dose warfarin, a


      clinically relevant comparator.  This superior


      protection was consistent across multiple


      subgroups.  These data support the efficacy of


      ximelagatran for the indication requested.


                The third indication considered today is


      the protection of patients from stroke and other


      thromboembolic complications of atrial


      fibrillation.  We have demonstrated that


      ximelagatran provides this protection, as well as


      does warfarin, across a broadly based patient




                Two independent pivotal trials--SPORTIF


      III, dosed, open-label in 23 countries in Europe


      and Asia, and SPORTIF V, conducted double-blind in




      North America--enrolled patients eligible for


      warfarin therapy according to existing treatment


      guidelines, that is, those with nonvalvular atrial


      fibrillation with at least one additional risk


      factor for stroke.  Each SPORTIF trial by itself


      studied more patients than all previous trials of


      stroke prevention in atrial fibrillation combined.


      Each trial compared 36 milligrams twice daily


      ximelagatran to dose-adjusted warfarin in


      preventing all strokes and systemic embolism, hard


      clinical endpoints in an intention-to-treat




                The choice of 36 milligrams came from


      several considerations.  There is no surrogate


      marker for stroke and systemic embolism, and both


      events are devastating.  Thus, we performed a


      dose-ranging study for safety and tolerability of


      ximelagatran 20, 40, and 60 milligrams in SPORTIF


      II, a 3-month Phase II atrial fibrillation study.


      While the numbers were small in that study,


      bleeding was most frequent with 60 milligrams and


      also the warfarin comparator, and less frequent




      with 20 or 40 milligrams ximelagatran.  We knew


      that 24 milligrams was effective in the Phase II


      European orthopedic surgery program and reasoned


      that any downside impact of potential additional


      strokes with 24 milligrams would be far worse than


      the bleeding seen with 36 milligrams in this


      nonsurgical context.  Using this educated judgment,


      we chose 36 milligrams in the Phase III atrial


      fibrillation program.


                Let's take a moment to consider the


      open-label nature of the SPORTIF III trial.  The


      majority of prior stroke prevention trials in


      atrial fibrillation also utilized an open-label


      format based on the difficulty of managing


      anticoagulation in blinded fashion.  SPORTIF III


      featured open-label dosing at sites, but also


      centrally randomized allocation and two additional


      levels of blinding:  blinded local assessment of


      primary endpoints by study-affiliated neurologists,


      and blinded independent central committee


      adjudication of all study endpoints.  SPORTIF V


      featured double-blind, double-dummy medication, and




      for patients receiving ximelagatran and placebo


      warfarin, sham INR values that mimicked those


      obtained during warfarin therapy.


                The established efficacy of warfarin


      precluded a placebo comparison.  Because warfarin


      is so efficacious, it is reasonable to establish


      ximelagatran efficacy in comparison to warfarin,


      and we did so using a noninferiority design.  In


      consultation with an executive steering committee


      and data safety monitoring board compose of leaders


      of prior stroke prevention trials and a


      statistician expert in noninferiority trials, we


      prespecified a 2-percent per year absolute


      noninferiority margin.  The choice of this margin


      has been questioned.  The choice of 2 percent arose


      partly from an expected 3.1 percent warfarin rate,


      but more importantly, from consideration of the


      clinically tolerable absolute difference in stroke


      rates considering warfarin's overall clinical


      profile.  A similar consideration drove designers


      of the SPAF III trial to power that trial to detect


      a 2-percent per year event rate with upper




      confidence bounds of 3 for a population at lower


      risk of stroke.  Even so, we prespecified a more


      conservative 2-percent upper confidence limit.  The


      point estimate of the difference in event rates


      needs to be much smaller than 2 percent for the


      worst case, that is, the upper confidence limit, to


      be less than 2.


                The strength of the 2-percent per year


      absolute margin resides in its clinical relevance,


      its prespecification, and that it is conservative.


                At screening, those patients already


      taking oral anticoagulants interrupted that therapy


      to decrease INR to 2 or less by the time of


      randomization, at which time patients received


      either warfarin or ximelagatran.  Each trial


      achieved a degree of warfarin control rarely found


      in routine clinical practice.  The warfarin-treated


      groups constituted formidable comparators for


      ximelagatran, particularly in SPORTIF V.  Samsa and


      colleagues found that most patients taking warfarin


      spend more than half the time on treatment outside


      the therapeutic range.  In SPORTIF V, only 15




      percent of patients did so.


                The ximelagatran- and warfarin-treated


      cohorts displayed nearly identical demographic


      profiles in each independent Phase III trial, seen


      here as pooled data.  Patients reflected well the


      elderly population of nonvalvular atrial


      fibrillation patients requiring anticoagulation for


      stroke prophylaxis, and the majority had impaired


      renal function.


                In SPORTIF III, warfarin, shown in gray,


      displayed an event rate of 2.3 percent per year


      compared to 1.6 percent per year with ximelagatran,


      shown in yellow.


                In SPORTIF V, the rates were 1.2 for


      warfarin, in gray, and 1.6 for ximelagatran, in




                Primary event rates with ximelagatran are


      nearly identical in SPORTIF III and SPORTIF V.  For


      warfarin, the rates fall within the range of event


      rates in previous trials, 0.6 to 4.1 percent per




                For comparison, the pooled rate in prior




      stroke trials for patients in this risk category


      taking placebo or aspirin was over 8 percent per


      year.  The difference in event rates in SPORTIF


      III, 0.66, favoring ximelagatran, had an upper


      confidence limit of 0.13, less than the


      prespecified 2-percent margin.  In SPORTIF V, the


      difference of 0.5 favoring warfarin had an upper


      bound of 1.03, also less than the prespecified


      2-percent per year margin.  Thus, each trial


      independently succeeded by satisfying the


      prespecified noninferiority criterion for the


      primary outcome.


                As expected, most of the events in this


      composite outcome were ischemic strokes.


      Hemorrhagic stroke and systemic embolism occurred


      more rarely and did not influence the primary


      endpoint substantially.


                Several sensitivity analyses confirmed the


      results of the primary analysis.  One such


      analysis, depicted here, included all-cause


      mortality in the primary endpoint at the suggestion


      of the agency.  SPORTIF III returned event rates of




      4.2 and 5.1 for a difference of 0.87 favoring


      ximelagatran, while SPORTIF V rates were nearly


      identical at 4.7 and 4.8.


                Adding all-cause mortality shifted each


      study's event rate difference point estimate to the


      left in favor of ximelagatran.


                Another sensitivity analysis, depicted


      here, used an on-treatment approach using the same


      endpoints and the same population, but not counting


      events that occurred after stopping study treatment


      for 30 continuous or 60 total days.  The upper


      bound of negative 0.18 indicates superiority in


      SPORTIF III.  The value of 1.2 in SPORTIF V


      indicates noninferiority to well-controlled




                For each trial, we also performed a paper


      comparison of ximelagatran to placebo by factoring


      in the results of the six prior stroke prevention


      trials.  We obtained original data from those


      trials to utilize the same endpoint events as in


      SPORTIF.  Demographics of patients in these trials


      were similar to those of SPORTIF patients.




                In these calculations, SPORTIF III and


      SPORTIF V separately demonstrated statistically


      significant risk reductions for ximelagatran


      relative to putative placebo, as did the pooled


      SPORTIF data.  Ximelagatran works as an


      anticoagulant in this population.


                As before, here we see differences in


      primary event rates according to demographic


      subgroups.  These pooled results reveal no


      discrepancies in any particular subpopulation,


      including the elderly, women, the obese, and those


      with poor renal function.


                In conclusion, for atrial fibrillation


      each of two trials independently met its objective,


      demonstrating that 36 milligrams of ximelagatran


      taken twice daily prevented stroke and systemic


      embolism to an extent similar to that of


      well-controlled anticoagulation with warfarin.


                For long-term secondary VTE prevention,


      the THRIVE III trial demonstrated that 24


      milligrams ximelagatran twice daily prevented VTE


      recurrence compared to placebo.




                And in total knee replacement surgery, the


      two independent EXULT trials showed that 36


      milligrams twice daily prevented VTE and all-cause


      mortality better than dose-adjusted warfarin.


                Based on five pivotal trial, each the


      largest in its field, involving more than 12,000


      patients, these data establish the effectiveness of


      ximelagatran as an oral anticoagulant in a variety


      of patient populations at high risk for




                Dr. Sunita Sheth will next address


      particular safety aspects of administration of


      ximelagatran for these indications.


                DR. BORER:  Thank you very much, Jay.


                Again, we'll take a minute to see if


      anyone has any issues that require clarification.


      Clearly, we are going to talk about or probably


      we're going to be talking about the selection of


      the delta for the noninferiority trial, but I don't


      want to get into that discussion now.  We have some


      extraordinary statistical fire power here between


      Tom on the committee and Lloyd Fisher and Jerry




      Faich sitting over there and the FDA statisticians.


      I think we'll wait on that until after all the


      presentations, including the FDA presentations,


      have been made.  But if we have any issues of fact


      that need to be clarified now, let's do it.




                DR. SACKNER-BERNSTEIN:  In the FDA


      briefing document, it points out that there were


      patients who were withdrawn from the study for whom


      there is not information about whether they


      underwent or suffered any events.  If that's


      correct, please clarify, because it looks as though


      from the study flow that that means in SPORTIF V as


      many as 15 percent of the patients we basically


      would not have any clinical outcomes data available


      from the point in time when they withdrew.  Is that




                DR. HORROW:  We followed up on all of our


      patients in the SPORTIF III and V trials to the


      greatest extent possible, and, in fact, after we


      were done with our follow-up, at the time of final


      closure, locking the database, we were left, out of




      7,922 patients, with only 63 patients about whom we


      were unsure of their final status.


                DR. SACKNER-BERNSTEIN:  So that would mean


      that the FDA briefing document is incorrect,


      because the FDA briefing document states that--and


      I'm looking at page 36 of the clinical review from


      Cardio/Renal Division.  It says in the first


      paragraph that patients that were discontinued from


      study medication and withdrew from study were not


      followed for primary efficacy endpoints or death.


      And then as you turn to page 45 with the patient


      disposition in SPORTIF V, it looks as though


      there's 300 study withdrawals from the ximelagatran


      group and 286 from the warfarin group.  So that


      means that about 15 percent would have incomplete


      clinical outcomes data, but you're saying there's


      only 63.


                So could you explain for us where the


      disparity should be settled?


                DR. HORROW:  It is conceivable that there


      is a misinterpretation of the term "study drug


      discontinuation" and "withdrawal from study."  More




      likely, the misunderstanding may accrue from the


      follow-up efforts that we made to ascertain the


      vital status of every patient in the SPORTIF




                We followed up on every patient, aside


      from the 63 that I just mentioned, and are


      confident in their vital status, knowing whether


      they were alive or dead, whether they had a stroke


      or not, in our database.


                DR. FLEMING:  Could I just clarify?  So I


      assume what you then did is you defined a date of


      data lock or closure where on that calendar date


      you wanted to follow all patients relative to their


      survival status and stroke status.  Are you saying


      then for all but 63 patients you knew their


      survival status and stroke status as of that


      calendar date for data lock?


                DR. HORROW:  Exactly, and that would be


      the data lock date for each respective


      trial--SPORTIF III and SPORTIF V.  That is correct.


                DR. BORER:  Steve?


                DR. NISSEN:  I want to make sure I




      understand how you maintained the blind,


      particularly in SPORTIF III.  Obviously with


      warfarin, you may require frequent dose adjustments


      and so on.  So in the open-label, particularly


      SPORTIF III, how did you maintain--in both trials,


      I'd like to understand the procedures that were


      undertaken.  I guess in the open-label trial there


      was no blinding, right?  The physicians and


      patients knew what they were receiving; is that




                DR. HORROW:  In the open-label trial?


                DR. NISSEN:  Yes.


                DR. HORROW:  It was open-label dosing, and


      so you are correct that the physicians and the


      patients knew the drug, and the evaluators, the


      neurologists locally, and the central adjudication


      committee were blinded and didn't know.


                DR. NISSEN:  Okay, I understand.  And


      SPORTIF V, then, how did you adjust warfarin and


      maintain the blind?  Explain to me how that was




                DR. HORROW:  It was quite tricky and




      involved quite a bit of work on the basis of the


      investigators and quite a burden for the patients.


      In SPORTIF V, all of the INR values were obtained


      in almost all cases by only two


      laboratories--that's an incredible degree of


      standardization for thromboplastin--either the


      centralized laboratory or a point-of-care machine


      that had standardized cards.


                In each case--well, for the point-of-care


      machine, a coded number was produced by the


      machine.  That was called in to a central


      randomization area, and that service then faxed to


      the site either the true INR value if the patient


      was really in the warfarin group or a shammed INR


      value if the patient was truly taking ximelagatran.


      So the site was unaware when it received the fax


      what group the patient was in.


                If the test was done at the centralized


      laboratory, then the centralized laboratory


      likewise sent the results to the IBRS site, the


      specialized service, which then, again, faxed


      either the shammed or the true INR value to the






                DR. NISSEN:  And then dose adjustments,


      how were those then made?  I mean, obviously some


      of the patients needed a dose adjustment, so what


      happened then?


                DR. HORROW:  Well, as you know, for


      ximelagatran or its placebo there were no dose


      adjustments.  But for warfarin or its placebo, each


      investigator adjusted the dose based on their usual


      practice considering the patient and the INR value


      or shammed value--they didn't know which it


      was--they'd received by fax.


                DR. NISSEN:  So there was no--it was all


      done per local physician practice.  There was no


      standard applied to how dose adjustments were made.


      Is that right?


                DR. HORROW:  That's correct.  We did not


      require all the investigators to adjust their


      patients' warfarin doses against some standard.


      This was to be a very real-world--as much as we


      could--type of adjustment in terms of warfarin or


      shammed dosing.




                DR. NISSEN:  And I assume the reason you


      didn't do that in SPORTIF III was that you just


      felt it was too difficult.


                DR. HORROW:  In SPORTIF III, the


      investigators were very uncomfortable with blinded


      anticoagulation testing and were unwilling to move


      forward in that regard.


                In SPORTIF V, our North American


      investigators embraced the randomization somewhat


      more willingly.


                DR. NISSEN:  So you tried to do SPORTIF


      III blinded but they wouldn't go along with it?  I


      don't understand exactly what happened.


                DR. HORROW:  It was not possible to get


      the investigators in SPORTIF III to move forward


      with the blinded testing and anticoagulation.


                DR. NISSEN:  You attempted it, and then


      they weren't able to comply.  Is that what




                DR. HORROW:  At an investigators meeting,


      there was--


                DR. NISSEN:  A rebellion.




                DR. HORROW:  There was no support.


                DR. BORER:  Okay.  Bill?  And then we have


      Tom and Alan and John.


                DR. HIATT:  A comment and a question.  In


      the knee replacement studies, you commented that


      you achieved a rapid increase in INR and that it


      was 2.4 at day 3.  And I just want to comment that,


      you know, there's an association between


      antithrombotic and anticoagulant effects of


      warfarin.  It takes 4 to 6 days for Factor II to be


      depleted, so that's a false sense of security


      around the measurement of the INR.  They're still


      not antithrombotic.


                So my question is:  If you take the


      three-quarters of patients at the end of that study


      who were, quote, therapeutic versus the one-quarter


      that were not, did you look at a subgroup analysis


      around difference in VTE rates at the end of that


      time?  Were the patients who were, in fact,


      therapeutic by that number equivalent in terms of


      VTE rates compared with the patients who were






                DR. HORROW:  My understanding of the


      question is did we perform a subgroup analysis near


      the end of the treatment interval regarding


      patients--or based on the actual INRs of the


      patients.  We do not have that analysis.


                DR. HIATT:  I think the speculation would


      be that the differences would be erased in those


      who were therapeutic, and a major difference


      between treatments would have been in those who


      were sub-therapeutic.  That was my question.


                DR. HORROW:  This is quite possible, and


      it's important to understand that the EXULT trials


      mimicked warfarin administration in the orthopedic


      surgery realm as it is currently practiced today in


      the United States.  And so it was a very relevant


      way to look at the effects.


                DR. BORER:  Tom?


                DR. PICKERING:  Can you tell us how the


      INR control rates in the SPORTIF trials compared


      with the same rates in the six


      warfarin-versus-placebo trials?


                DR. HORROW:  The INR rates in the six




      index trials had somewhat of a spread, as would be


      expected, and it's actually possible to see that as


      the INR rates are better in some of those trials,


      so are the results in terms of the decrease in the


      warfarin event rate.  And our results for INR


      control were really quite in the middle, 2.5,


      2.4--could we have the previous slide, please?  I


      can show you some data on them.


                This would be for SPORTIF V, summary


      statistics.  Please note in the middle column


      labeled ximelagatran, we are looking at shammed


      values, and you will note that we have 2.5 at 3


      months for ximelagatran and 2.4 for warfarin, at 12


      months similarly, at 24 months similarly--right in


      the middle of the desired interval.  And, of


      course, the other thing that you might note here is


      that there is a threshold of 4.0 for the shammed


      values to ensure that no shammed value ended up


      putting a patient unnecessarily in the hospital


      because of an elevated shammed INR.


                Nevertheless, as you can see by the ranges


      here, it's quite clear that the investigators would




      be unable to determine whether a patient were in


      one group or the other.


                DR. PICKERING:  That really wasn't my


      question.  I was asking if there are comparable


      data for the six warfarin placebo trials.


                DR. HORROW:  I don't have those data


      available to show you at this time.


                DR. BORER:  I think they're in one of our


      two books, Tom.


                DR. HORROW:  I believe they may be in the


      briefing document.


                DR. BORER:  If I remember correctly, they


      do show a fairly wide range, as you might expect,


      but we can get those data.


                Okay.  Alan?


                DR. HIRSCH:  I have two questions.  One is


      to follow up Steve's question regarding the SPORTIF


      III blinding.  I just always believe it's terribly


      important to have blinding as a component of major


      pivotal trials, acknowledging that lack of blind


      can really alter outcomes in unexpected ways.  So I


      want to just run this through one more time.




                Pitying the investigators that would not


      go along with your request, the patients knew their


      study assignment, correct?


                DR. HORROW:  In SPORTIF III.


                DR. HIRSCH:  In SPORTIF III.


                DR. HORROW:  That's correct.


                DR. HIRSCH:  The physicians--


                DR. HORROW:  The patients knew their


      assignment, as did the principal investigators.


                DR. HIRSCH:  And coordinators.


                DR. HORROW:  That's correct.


                DR. HIRSCH:  So how would we have any


      confidence that the adjudicating neurologist would


      have any blind maintained at all?


                DR. HORROW:  Well--


                DR. HIRSCH:  I worry.


                DR. HORROW:  Your point is well taken that


      that cannot be assured with certainty.  We can say


      that there were efforts made to make sure that the


      neurologist was not told on purpose the assignment


      of the patient, and we know also that all members


      of the central adjudication committee, which




      evaluated all the endpoint events upon which the


      results are based, did so in a totally blinded




                DR. HIRSCH:  I guess if there was


      concordance between those two groups, I'm somewhat




                Let me come back with a follow-up question


      for EXULT, if I could.  The data that we have


      demonstrates benefits of ximelagatran versus


      Coumadin preventing DVT in this population at risk


      after total knee replacement.  And as we'll discuss


      later, most of that data is regarding distal DVT,


      which I do care about.  But in the database, do we


      have any evidence, quality-of-life measurements,


      girths, anything that demonstrates a clinically


      relevant effect for the patient?  In other words,


      in the absence of venographic surveillance, would


      the patient know there was a difference in outcome?


                DR. HORROW:  I'd like to ask Dr. Scott


      Berkowitz, who is the medical director for that


      particular trial, to address that issue.  Dr.






                DR. BERKOWITZ:  Hi.  Scott Berkowitz,


      AstraZeneca.  There was not any type of


      quality-of-life assessment in this short-term


      trial.  The symptomatic events were collected as


      well, including distal, proximal, and PEs.  They


      were low, as they are in TKR trials and did not see


      a difference, a statistical difference.


                DR. BORER:  There was a question I was


      going to hold until the end, but it seems to be


      relevant right here in view of Alan's point.  You


      probably have a back-up slide, and Alan just


      suggested that he probably has the data off the top


      of his head.  But can you tell us, among people


      historically from older trials where data would be


      available who have asymptomatic distal DVT and who


      aren't treated, what's the risk of subsequent


      thromboembolic events during some follow-up period?


                DR. BERKOWITZ:  Well, we don't have the


      greatest data on that, unfortunately, in the


      literature.  What we know is that 10 to 20 percent,


      depending on what you're readings--there are only


      three or four studies--do propagate from distal to




      proximal.  We know about 5 percent propagate to PE.


      We don't know the actual recurrence rate of what


      further DVTs would be after, say, 6 months.  We do


      also know that post-thrombotic syndrome occurs in 5


      percent of patients in 2 to 7 years after total


      knee replacement.  Those are the real data that we


      have.  Not an area well studied.


                DR. BORER:  Jonathan, and then Steve.


                DR. SACKNER-BERNSTEIN:  I noticed that in


      the trials for the study flow of patients in


      several of your trials, including THRIVE, both


      EXULTs, and SPORTIF V, that there is a number of


      patients listed as being enrolled and then a second


      number of patients listed as being randomized.  And


      there's very little information in either the FDA


      or the sponsor's documents about what happened to


      those patients.  So I'm wondering if you could


      describe it because in each of the cases you're


      looking at probably in the range of 10 percent of


      the patients who are enrolled that don't make it to




                DR. BERKOWITZ:  Maybe I'll first try to




      answer for the EXULT and THRIVE, and then ask Dr.


      Horrow for atrial fibrillation.


                For the EXULT trials, patients were


      enrolled, meaning that they were seen as an


      outpatient up to a month before the procedure, and


      then would come into the hospital, and if they had


      the surgery of interest, which was primary total


      knee replacement, then would be randomized,


      assuming they went through the eligibility


      criteria.  The most common reason patients wouldn't


      go from enrollment to randomization is that either


      the--there were two:  one, that the surgery was


      cancelled, and then the patient wasn't rescheduled


      for the procedure--excuse me, for the study, but


      did do the procedure; the other was that with these


      trials rapidly enrolling, we had many people lined


      up but then the study--we reached our enrollment.


      Those were the two major causes.


                For the THRIVE study, these were patients


      who had acute events for 6 months treated acute DVT


      and then went on to a 6-month--either placebo or to


      ximelagatran 24-milligram arm, and most of these in




      terms of just taking a look here--I can just show


      you what we've got in terms of that.  In terms of


      the ones that were not randomized, there were 123


      of those patients, and most of this turned out to


      be eligibility not fulfilled or withdrawn consent.


      And that is a common thing that patients might


      think more about the study if they want to


      participate in such a long-term--and then I could


      turn it over to Dr. Horrow.


                DR. HORROW:  In the SPORTIF trials, the


      major reason why patients were enrolled but not


      randomized was because of the failure of an


      eligibility criterion; in particular, the major one


      was the ability to achieve two electrocardiograms


      demonstrating atrial fibrillation in the manner


      specified.  And as a result, the principal


      investigators did not enroll a number of the


      patients whom they at first thought were good




                DR. SACKNER-BERNSTEIN:  Can I just follow


      up?  One quick point in follow-up.  In the patients


      enrolled in SPORTIF where many of them were coming




      off the vitamin K antagonist, how many--even if it


      was a minority, how many of those patients had some


      sort of clinical event that led them not to be




                DR. HORROW:  I understand your interest is


      in seeing what happens to the patients who came off


      of vitamin K antagonist in the enrollment period,


      did they happen to have events.  I believe that we


      have some data on that, although I can't say for


      sure that all of these did not enroll.  They may


      have had an event after enrollment.  If you'll just


      give me a moment, I'll see if we can find these




                Yes, thank you.  Here we see the number of


      patients with primary events who had an event


      within 30 days of discontinuing study drug, and


      this would be either the ximelagatran or the


      warfarin group.  And this would be during the


      course of the trial.  As you can see, there's not


      much difference between the two groups.


                I think this may address the question that


      you're getting at, which is what happens when




      patients discontinue their anticoagulant.


                DR. SACKNER-BERNSTEIN:  Well, actually, I


      find that reassuring, that information, but really


      what I was getting at was the impact of the


      strategy that would be proposed based on the study,


      which is you have a patient who's on long-term


      warfarin and you're going to convert them


      potentially to a new agent.  There's a period that


      would be followed where there's a transition, and


      I'd like to know if that transition period is a


      period that could be associated with risk as well.


                DR. HORROW:  I understand better.  Thank


      you.  Here are some data from SPORTIF III looking


      at primary events within 7 days of randomization.


      There were three patients who had a primary event


      in the SPORTIF III trial within 7 days of


      randomization, and, of course, the patients taking


      VKA--all patients had to stop their VKA in order to


      begin randomization.  And there were two events in


      the warfarin group and one in the ximelagatran




                DR. BORER:  Steve?




                DR. NISSEN:  I want to come back to the


      blinding issue again, and we've been dancing around


      it so let me just come to the point.


                Something extraordinary happened in




      calculate a hazard ratio of 1.39, 1.40 that's in


      favor of ximelagatran.  And in SPORTIF V, the


      hazard ratio is 1.35 in favor of warfarin.  And so


      you have almost a completely opposite effect on the


      point estimates, which, you know, is really unusual


      when you consider the similarity of the trials.


                So we're al trying--we're all sitting here


      looking at the briefing document, and we're trying


      to figure out what could possibly have happened


      here so that, you know--I mean, there's essentially


      a 39-percent greater risk for warfarin in SPORTIF


      III and a 39-percent greater risk for ximelagatran


      in SPORTIF V.  And the only big difference in the


      two trials is that one was blinded and one wasn't.


                And so most rational people who look at


      that would say, well, we're going to believe the


      blinded results, we're not going to believe the




      unblinded results.  And so this is a real


      credibility issue, and I think we might as well


      just put it on the table and get your reaction to




                DR. HORROW:  In fact, there are many


      differences between SPORTIF III and SPORTIF V that


      are confounded with the open-label and double-blind


      nature of those two trials.  The first and foremost


      is geography, namely, that one study was conducted


      in Europe and Asia and the other in North America,


      and practice issues may pertain.


                Secondly, although SPORTIF V patients more


      often had hypertension, their blood pressures were


      6 mm mercury lower, on average, than patients in




                And, third, there was an artificially


      intense control of INRs in SPORTIF V relative to


      SPORTIF III, because in SPORTIF III there were over


      270 clinical laboratories conducting INR


      measurements, but there were essentially two in


      SPORTIF V, achieving some kind of standardization


      that is difficult to quantify.




                Another aspect that is important to


      consider is that the ximelagatran rates were


      identical in the two trials.  And in the warfarin


      trials--I'm sorry, in the two trials, the


      ximelagatran rates were identical, about 1.6.  The


      warfarin rates appear disparate.  But those rates


      are actually within the range of rates that are


      seen in prior stroke prevention trials.


                What we may be looking at here is another


      manifestation of the variability of warfarin.  This


      slide shows in yellow the warfarin rates from the


      six index trials, in orange the two rates from the


      SPORTIF trials, and in dark brown the meta analysis


      rate for the trials in yellow.  And as you can see,


      the SPORTIF rates are within the range of the


      warfarin rates from the previous trials.


                I hope that gives some perspective.


                DR. FLEMING:  Could you put that slide up


      again?  Can I follow up?


                DR. BORER:  Sure.  Let me just put some


      ground rules here, though.  Steve has highlighted


      what will be one of the key issues for discussion




      later on, and rather than get into it in great


      detail here and get bogged down for the next hour,


      perhaps we can deal only with issues of fact, and


      then we'll get into the evaluation of those facts a


      little bit later.


                But with that in mind, go ahead, Tom.


                DR. FLEMING:  If you could put that slide


      up, I just think for clarification, I don't think


      that the point you just raised really answered


      Steve's question.  Steve's question had more to do


      with the heterogeneity in the relative risk


      estimate across to pivotal studies.  This is


      getting at the heterogeneity of the control arm


      event rates across trials.  And, in fact, those are


      different phenomenon.  This really gets at the


      unreliability of noninferiority comparisons because


      of this tremendous heterogeneity, which is a


      separate issue.


                While I have the mike, could I ask a


      question that I had in mind?  That is, one of the


      things that's always concerned me in trials with


      venograms is that we end up with a lot of missing




      data, far more than what this committee would be


      used to accepting in a manner to maintain integrity


      of randomization.  I think you had 20 and 15


      percent, respectively, missing the outcome


      assessments in EXULT A and EXULT B.


                With that in mind, and also wanting to


      really focus on what are not surrogates but true


      clinical endpoints, endpoints that reflect tangible


      benefit to patients, I struggle to look for what


      are those measures that are really tangible that


      are measured uniformly in patients.  Could you show


      Slide CE-19 as we look at EXULT A and B?  Two of


      these measures are pulmonary embolism and death


      that should be assessed, I'm assuming, and


      available in all patients.  Your survival figures


      here reflect, if I pool here, five deaths against


      three.  The agency on page 26 of their briefing


      document has ten against four, so you're missing


      five deaths in the Exanta arm and one in the


      warfarin arm.  Could you clarify that discrepancy?


                DR. HORROW:  If I may first address the


      issue of the heterogeneity, then we can go on to




      the issue with EXULT.


                If I'm not mistaken, you're referring, in


      terms of the heterogeneity in SPORTIF, to what may


      be called a study by treatment interaction, the


      difference in sampling and getting one set versus


      the other.  And I think it's important to


      understand that in each case, noninferiority was


      satisfied; that is, looking at the data just in


      those terms and how those numbers are sorted does


      not take into account the noninferiority design of


      the trials and that the success is determined by


      whether or not it meets the noninferiority




                The heterogeneity result which we've


      looked at is not robust to sensitivity analyses


      like the primary results are robust.  So, for


      example, if one looks at primary events plus


      all-cause mortality, which was an endpoint


      suggested by the agency, the heterogeneity


      disappears and the p value is 0.23.  And if you


      look at other prespecified outcomes, such as major


      bleeding, there's no suggestion of disparity there.




      The heterogeneity p value is 0.81.  For total


      bleeding it's 0.275.


                And so we view the idea of disparate


      results in the two trials with some suspicion and


      think that we need to be very careful how we


      interpret those primary results in terms of being


      disparate or the same.  We view them as sampling


      from the same pool and getting two separate results


      and that the best estimate of the data comes from




                I'd like now to--


                DR. FLEMING:  Given that you didn't answer


      my question and you provided a different answer,


      let me respond to the answer you just gave.  The


      question that Steve asked is why was there such


      heterogeneity in relative risk estimates.  The


      answer that you gave was there's a lot of


      heterogeneity in the control arm, in the warfarin


      rates across trials.  Logically, I would assume


      that if you're saying when the warfarin rate in


      truth is different across trials, we should expect


      a different treatment effect, it really makes me




      worry about doing a noninferiority trial where you


      have to rely on historical evidence.


                Could you answer, though, the question


      that I'd asked here about the discrepancy between


      your data here and the FDA briefing document?


                DR. HORROW:  I'd like to ask Dr. Scott


      Berkowitz, who was the medical person for this


      particular trial, to address this issue.


                DR. BERKOWITZ:  Yes, Scott Berkowitz,


      AstraZeneca.  I just wanted to say in terms of the


      venography rate--I have the data to show you, but


      in terms of venograph, these two trials had the


      highest adequacy of evaluability ever done in


      clinical trials for pivotal purposes.


                DR. FLEMING:  That may be, and yet the


      reality is we're still lacking 15 to 20 percent of


      our randomization cohort, and we no longer are


      assured of integrity of randomization.  So could I


      get the answer to my question?


                DR. BERKOWITZ:  So for what you saw, those


      data that you saw in the briefing packet were for


      the overall study, so you can see it's ten and




      four, but I'm going to--could I have the next


      slide?--show you the breakout for treatment, which


      is the primary endpoint--


                DR. FLEMING:  So, in fact, what I do want


      is the entire study, ten and four.  So is the


      clarification CE-19, then--


                DR. BERKOWITZ:  Could we go back?


                DR. FLEMING:  Then the reason CE-19 is


      leaving out the five deaths and one death is that


      those occurred in the non-80, 85 percent?


                DR. BERKOWITZ:  The deaths--I'm sorry.


      Say that again?  I'm sorry.


                DR. FLEMING:  What is the reason that your


      slide here leaves out five deaths and one death?


                DR. BERKOWITZ:  That slide showed the


      primary endpoint which included the treatment


      period of day 7 to 12 days as opposed to the


      overall, which showed only this study and the next


      one, if you want to see the breakout.


                DR. FLEMING:  Good.  And so that is--could


      you show it again?


                DR. BERKOWITZ:  Oh, yes.  Can I see the




      next one?  Thank you.  We want to see now the


      breakout between the treatment--


                DR. FLEMING:  So that the total deaths are


      as here, they are as there in the FDA briefing


      document, ten against four in the wrong direction.


      And pulmonary embolism is, according to the FDA


      briefing document, four against five as reported by


      the FDA.


                DR. BERKOWITZ:  Yes.  Well, you can see


      down--for the treatment period, as you can see,


      there was one in the ximelagatran 36 group and none


      in warfarin.  During the follow-up period, there


      were four in the ximelagatran group and zero in




                DR. FLEMING:  And so in an ITT analysis


      that does include all patients and focuses on,


      among the most clinically relevant endpoints, death


      and pulmonary embolism, it appears that there are


      actually numerically an excess of events in the


      Exanta group.  By my count there are 15 events


      against 10 events, and that's your numbers as well.


      Is that correct?




                DR. BERKOWITZ:  Well, except that the


      numbers that you're seeing in follow-up are after


      patients are off treatment but they get seen in 4


      to 6 weeks.


                DR. FLEMING:  I want ITT, and that's what


      it looks like.  Is that correct?  It's 15 against


      10 in the wrong direction?  Just is the FDA summary


      correct on page 26?


                DR. BERKOWITZ:  Yes.


                DR. FLEMING:  And one other quick


      question, if I could.  Again, wanting to try to


      focus on an ITT of a critical endpoint, all-cause


      mortality, in THRIVE III could you show us the ITT


      summary?  This is the placebo-controlled trial


      where we see a substantial efficacy result on the


      symptomatic endpoint.  Could you show us the ITT of


      the survival curves for that trial?


                DR. BERKOWITZ:  I'm not certain--did you


      want to see the slide that we showed for the


      original presentation?


                DR. FLEMING:  I believe it's corresponding


      to the page 7, Figure 1 in your briefing document.




                DR. BERKOWITZ:  Let me bring that up.  I


      just want to be sure it's the same one that we saw.


                DR. FLEMING:  The one that I am in


      particular looking for here, because that figure


      includes all the data from all the trials, is in


      particular THRIVE III with ITT analysis of


      mortality over the time frame that you followed


      these patients.


                DR. BERKOWITZ:  Yes, okay, and that's what


      we were--yes, I'm sorry.  So here you go.  This is


      the slide.


                DR. FLEMING:  Mortality.


                DR. BERKOWITZ:  Yes.


                DR. FLEMING:  All-cause, ITT.


                DR. BERKOWITZ:  I'm sorry.  I still don't


      understand what you--just the mortality slide?


                DR. FLEMING:  Yes, as you have in Figure 1


      of your briefing document.


                DR. BORER:  You wanted to see only for


      THRIVE, or you wanted to for the--


                DR. FLEMING:  Either way, if you--okay.


                DR. BERKOWITZ:  This is the slide in the




      briefing document that you're speaking of with all


      the mortality.


                DR. FLEMING:  Okay.  And could you--so the


      THRIVE is, in fact, the--


                DR. BERKOWITZ:  I'm sorry, yes, the THRIVE


      is the lowest curve there, the ximelagatran versus


      placebo in the lowest group.


                DR. FLEMING:  And so essentially, while


      I'm focusing on THRIVE, the evidence here would


      suggest, even in placebo-controlled comparisons,


      there's strong suggestion of no differences in




                DR. BERKOWITZ:  Well, I mean, they're


      lower with the ximelagatran group, but not a strong




                DR. FLEMING:  I'm sorry.  I don't--the


      curves look overlapping in the THRIVE III, and in


      the other studies they are very overlapping as




                DR. BORER:  Steve?


                DR. NISSEN:  Yes, I just had one more


      question, still trying to probe to understand the




      differences between SPORTIF III and SPORTIF V.


      Could you show us the INR values, that is, the


      degree of anticoagulation control in SPORTIF III


      and SPORTIF V for the warfarin arms.


                DR. BERKOWITZ:  I'll ask Dr. Jay Horrow to


      present that.


                DR. HORROW:  I'm sorry.  I missed the last


      two words in--


                DR. NISSEN:  Yes, I just want to see in


      the warfarin arm of the trials, I want to see what


      the INRs looked like in SPORTIF III and SPORTIF V.


                DR. HORROW:  Okay.  I believe these data


      will address your question.  There were almost


      100,000 different INR values, and this summary


      perhaps helps.  Here we have SPORTIF III and


      SPORTIF V and the percentage of time in specific




                DR. NISSEN:  It does.


                DR. HORROW:  Okay.  Thank you.


                DR. BORER:  A final question of fact,




                DR. SACKNER-BERNSTEIN:  I think the key




      thing is that all of the slides that show ITT are


      not true ITT analyses.  It's not just THRIVE.  It's


      THRIVE and EXULT, and they list that in the


      briefing document.  There are a lot of numbers


      where those are different, so we should just


      interpret it that way.


                DR. FLEMING:  It was part of the reason


      for my asking the question.  I wanted to get a


      verification that we were being shown, for


      endpoints such as mortality, a true ITT.  And I


      understand that they're telling us they are showing


      us a true ITT where you have uniform follow-up


      through a given calendar date at which the study


      data freeze would have occurred, and you would have


      complete follow-up on mortality for all patients.


      Is that what that Figure 1 showed?


                DR. HORROW:  Yes.


                DR. BORER:  Okay.  Thank you, Jay.


                DR. HORROW:  May I introduce Dr. Sunita


      Sheth, who will discuss particular aspects of


      safety for ximelagatran.


                DR. SHETH:  Good morning.  I'm Sunita




      Sheth, Senior Director of Clinical Research at




                You've just seen the efficacy data


      supporting the benefit of ximelagatran as an oral


      anticoagulant.  I'll now review the clinical safety


      date.  The analysis comes from a large data set


      with more than 30,000 subjects, many of the


      patients involved having serious underlying disease


      and receiving multiple drug therapy.


                First, I'll discuss by indication the


      adverse events and bleeding profiles.  Efficacy for


      any anticoagulant is balanced by risk of bleeding.


      Indeed, bleeding and the prevention of thrombosis


      derive from the same action of drug.  That's why


      bleeding was a prespecified endpoint in the pivotal


      trials.  And major bleeding was adjudicated in a


      blinded fashion in all Phase II and Phase III


      trials.  Then I'll focus on two specific topics:


      myocardial ischemic events and the hepatic


      findings.  Finally, I'll conclude with a review of


      overall mortality and summarize the key points for


      each indication.




                It may help if I display how we've


      organized the large data set.  It divides logically


      into three groups:  Phase I, surgical, and


      nonsurgical populations.  The Phase I population,


      composed primarily of healthy volunteers dosed for


      up to 8 days, didn't present any safety signals.


      Surgical patients, mostly from the orthopedic


      studies with dosing up to 12 days, have different


      safety issues, in particular, perioperative


      bleeding, and so they are reviewed as a separate




                The nonsurgical population primarily


      received drug for more than 35 days and provides


      the core safety evaluation of long-term dosing,


      with exposure up to 4 years.  Each population pool


      is large, allowing a detailed assessment of safety


      in each case.


                I will first review the safety for the


      surgical indication.  The North American surgical


      population has been termed "the warfarin comparison


      pool" and provides the safety data for the


      indication under consideration today, with




      post-operative dosing of either oral ximelagatran


      or warfarin after total knee replacement surgery.


      This pool includes data from three Phase III


      trials:  the two EXULT trials as well as an earlier


      study evaluating 24 milligrams versus warfarin.


      Overall, it includes 5,236 patients.


                In all graphs, ximelagatran will be shown


      in a shade of orange and the comparator in gray.


                Here's the summary of adverse events for


      the surgical pool.  Both treatment groups showed a


      similar frequency and type of adverse events.


      There didn't appear to be any dose response


      comparing the 24- and 36-milligram doses.  We can


      look more closely at the EXULT trials where both


      major and minor bleeding events underwent


      independent adjudication.  Rates of major bleeding,


      shown at the bottom of each bar, were 1 percent or


      less in all treatment groups, with no statistically


      significant differences for major bleeding alone or


      for the combination of major and minor bleeding,


      for which respective p values are shown.


                When you look at the data for the proposed




      36-milligram dose, there wasn't a difference in


      surgical outcome parameters, such as wound hematoma


      or intra-articular bleeding.  Additionally, the


      proportion of patients receiving transfusion and


      the volume of transfusion were similar in each




                Now, let me turn to the nonsurgical


      patients who comprise the long-term dosing group.


      This group is called the long-term exposure or LTE


      pool, with patients from all the Phase II and Phase


      III studies conducted so far involving dosing


      beyond a month's duration.  In addition to patients


      from the atrial fibrillation and venous


      thromboembolic secondary prevention indications,


      we've included data from two other disease areas


      where significant trials have been conducted,


      patients undergoing initial 6-month treatment for a


      venous thromboembolic event and patients post-acute


      coronary syndromes.  The overall ximelagatran


      exposure is substantial, a total of 6,768 patient


      years, with a median exposure of 370 days.


                Across this population, doses between 20




      and 60 milligrams have been used, although the


      majority of patients, 75 percent of them, received


      36 milligrams twice daily.  The comparator group


      includes both placebo as well as warfarin and is


      termed "the comparators' group."  In this group, 20


      percent of patients received placebo.


                I'll now comment on the different


      indication pools.


                In the VTE extended prophylaxis pool, both


      ximelagatran and placebo groups demonstrate similar


      frequency and types of adverse events.  The


      incidence of serious adverse events and


      discontinuations was actually lower in the


      ximelagatran group compared to placebo.


                In the same group, major bleeding occurred


      rarely, affecting six patients in the ximelagatran


      group and five patients in the placebo group.


      Ximelagatran and placebo groups also did not differ


      with respect to major or minor bleeding events.


                In the atrial fibrillation pool, the same


      frequency and types of adverse events were recorded


      in both the ximelagatran and warfarin groups. 




      Discontinuations were higher in the ximelagatran


      group, not because of symptoms but mainly due to a


      protocol-mandated discontinuation for ALT


      elevation.  I'll discuss this in detail shortly.


                In the atrial fib population, the rates of


      major bleeding with ximelagatran did not differ


      from those with warfarin.  Minor bleeding events


      occurred quite often in these trials and for that


      reason did not undergo adjudication.  Here we see


      the event rates for patients with one or more major


      or minor bleeding events.  Total bleeding occurred


      significantly less often with ximelagatran than


      with warfarin, with a p value of less than 0.001.


                Overall, with regard to adverse events and


      bleeding, ximelagatran compared to well-controlled


      warfarin following total knee replacement surgery,


      compared to placebo and extended secondary


      prophylaxis of VTE, and compared to warfarin in


      atrial fibrillation patients demonstrated no


      important differences in adverse events, bleeding


      profile, or the safety profile of the 24- and


      36-milligram doses.  In addition, a detailed




      subgroup analysis for bleeding supports the


      proposed fixed-dose approach for all types of


      patients studied.


                I'll now review two special safety topics,


      coronary artery disease and the hepatic findings.


      First let's address the coronary artery disease




                The agency has noted a possible imbalance


      in the frequency of myocardial infarctions.  Shown


      here is Table 12 from the FDA briefing document.


      The events shown here are investigator-reported


      events.  Note that the absolute number of


      myocardial infarctions observed in the EXULT trials


      was small, and there appears in a post-hoc analysis


      to be a significant difference with a p value of


      0.049.  However, this difference is driven by a


      single trial, EXULT A.  Furthermore, an analysis of


      other coronary artery disease events failed to


      reveal any significant difference.


                FDA Table 40 shows investigator-reported


      coronary adverse events from selected trials from


      the long-term pool.  This analysis suggested an




      increased frequency of total coronary adverse


      events in the VTE treatment population.  When VTE


      treatment and extended prophylaxis are evaluated I


      a post-hoc pooling and analysis, the p value is


      significant for both myocardial infarctions and


      other coronary artery disease events.  However,


      this finding was not observed in the much larger


      atrial fibrillation pool.  In addition, the trial


      in acute post-coronary syndromes where benefit was


      demonstrated is not included in this analysis.  In


      fact, when all three groups are pooled, no


      significant difference is observed for either


      myocardial infarctions or other coronary events.


                In addition to the investigator-reported


      events, the SPORTIF trials in atrial fibrillation


      with an active comparator, warfarin, and the ESTEEM


      trial in the post-ACS setting versus placebo


      provided an independent and objective assessment of


      myocardial infarctions.  In fact, adjudicated


      events in these trials represent over 90 percent of


      all MIs across the program.  Here, evaluation of


      the SPORTIF trials demonstrated an identical




      incidence while the ESTEEM trial demonstrated an


      actual reduction in myocardial infarctions.


                It is also relevant in this context that


      across the whole program we have no evidence of any


      rebound effects producing MIs after ximelagatran


      treatment was stopped.


                So with regard to coronary artery disease


      adverse events, while a concern was raised


      regarding a potential imbalance in events, a more


      comprehensive analysis focusing on both


      investigator-reported and objectively assessed


      events fails to identify an increased risk.


                I now want to turn to the unexpected


      results, the hepatic findings, and present a


      detailed review.


                We've taken the findings very seriously,


      and from the large database individual case


      analysis and consultation with hepatic experts,


      we've produced a thorough assessment.  I'll first


      review the laboratory findings followed by the


      adverse event data.


                Preclinical toxicology and the Phase I




      studies did not demonstrate any hepatic safety


      issue.  The surgical studies with up to 12 days of


      dosing didn't show any hepatic changes with


      ximelagatran, just the well-recognized enzyme


      elevation seen with heparin.  In the first Phase II


      long-term dosing study with ximelagatran in the


      atrial fibrillation patients, a signal of an


      asymptomatic increase in ALT greater than 3 times


      the upper limit of normal was noted.  Therefore,


      the standard laboratory testing that was being


      performed early in the development program was


      increased in the Phase III studies.


                The liver function testing panel consisted


      of alanine amino transferase, or ALT; aspartate


      aminotransferase, or AST; alkaline phosphatase, and


      total bilirubin.  These tests were performed


      monthly for the first 6 months of exposure, then


      every 2 months up to one year, and then quarterly.


      In addition, weekly testing and discontinuation


      criteria were defined.  These criteria were


      strengthened after one case of biopsy-documented


      hepatic necrosis.




                As mentioned, there was no increase in ALT


      greater than 3 times the upper limit of normal in


      ximelagatran patients undergoing total knee


      replacement compared to warfarin during treatment.


      At the 4- to 6-week follow-up, there were eight


      patients in the ximelagatran group and three in the


      warfarin group that developed an increase in ALT.


      In general, these increases occurred 3 weeks after


      discontinuation of drug.  It's important to note


      that two patients with the transaminase elevation


      in follow-up in the ximelagatran group had received


      low-molecular-weight heparin.  The ALT elevation in


      all patients but one in each group is documented as


      resolved.  We believe that patients undergoing


      orthopedic surgery with short-term dosing of


      ximelagatran are not at an increased risk of ALT


      elevations or liver injury.


                Now, let me summarize the incidence of


      enzyme elevations of the long-term exposure pool.


      The incident of ALT greater than 3 times the upper


      limit of normal was 7.9 percent for ximelagatran


      compared with 1.2 percent for comparators.  It's of




      interest to note that there was no difference


      between groups for isolated elevations of


      bilirubin.  The vast majority of these enzyme


      elevations were asymptomatic.


                Our experience shows that the time


      signature for ALT elevation follows a consistent


      pattern.  This graph depicts the number of patients


      with first ALT greater than 3 times the upper limit


      of normal over time.  The y axis represents the


      cumulative risk of an ALT greater than 3 times the


      upper limit of normal and the x axis time in


      months.  As can be seen, the occurrence increases


      above background rates after 1 month and approaches


      background rates after 6 months.  Ninety-three


      percent were detected during the first 6 months,


      and 98 percent within the first 12 months.


                I now want to turn to the disposition of


      patients with an ALT increase.  Of the 546 patients


      in the ximelagatran group that had an increase to


      greater than 3 times the upper limit normal, 46


      percent of patients continued to treatment and


      completed the study.  The other 54 percent




      discontinued study drug.  Overall, 96 percent of


      ximelagatran-treated patients returned to less than


      or equal to 2 times the upper limit of normal ALT,


      regardless of continuation or discontinuation of


      drug.  Of the 74 patients in the comparator group,


      31 percent continued treatment, and the other 69


      percent discontinued treatment.  Overall, 93


      percent of comparator-treated patients recovered.


                The algorithm allows continuation of


      treatment for mild and transient increases on drug.


      These data demonstrate the reversibility of the ALT




                Patients who continued drug recovered by a


      median of 28 days, and those who discontinued drug


      by a median of 40 days.  Eighteen patients were


      rechallenged early in the program.  Only two


      patients had a subsequent ALT rise.  One pt with a


      peak ALT of 10 times the upper limit of normal was


      rechallenged after 65 days and did not have a


      repeat elevation until 2 months later.  The second


      peak was at 3 times the upper limit of normal, and


      the drug was discontinued.




                The second patient did not have a true


      rechallenge, but had multiple episodes above 3


      times the upper limit of normal, but overall


      recovered with continuation of the drug.  There was


      no evidence in these or any other patients for an


      immunoallergic response.


                Hepatic experts that we consulted


      suggested that the elevation of ALT greater than 3


      times the upper limit of normal and clinical


      jaundice, in the absence of an alternative


      diagnosis, can be considered a signal of severe


      hepatic injury.  We selected a more conservative


      definition to standardize the levels and timing and


      included cases with ALT greater than 3 times the


      upper limit of normal and bilirubin greater than 2


      times the upper limit of normal, the latter


      occurring within one month of the ALT rise.


                A total of 37 patients, or 0.53 percent,


      in the ximelagatran group had this concurrent


      elevation of ALT and bilirubin, compared with five


      patients in the comparators' group, with an


      incidence of 0.08 percent.




                Please note that one additional case has


      been included in this analysis at the request of


      the FDA.  We had fully documented this case


      involving a fatal GI bleed in the submission and


      had also highlighted it as a case of interest in


      the safety review.


                I'll now review the outcome in patients


      with a concurrent increase in both ALT and


      bilirubin.  Confounding diagnoses were noted in 25


      of the 37 patients on ximelagatran.  Seven patients


      in the subset died of unrelated causes.  Twelve


      patients did not have an alternative diagnosis for


      the enzyme elevation.  Of these 12, two died with a


      GI bleeding event and will be discussed shortly.


      The ALT and bilirubin in all other patients


      recovered.  Of the five cases in the comparator


      group, four had an alternative diagnosis, and only


      one had an unexplained increase.  Two patients died


      from pancreatic cancer.  The other patients




                We have been investigating a possible


      mechanism for the hepatic changes, but so far this




      has not been elucidated.  Preclinical studies


      evaluating reactive metabolites, mitochondrial


      dysfunction, and protein binding have not been


      revealing.  There is no evidence for involvement of


      the P450 system.  The asymptomatic and


      nonprogressive pattern of ALT increase has been


      noted with other drugs, including tacrine, INH,


      amiodarone, among others.


                We wanted to understand if there's a


      subgroup that's at increased risk.  Because the


      number of patients with concomitant ALT and


      bilirubin is so low, this analysis was performed on


      the occurrence of ALT greater than 3 times the


      upper limit of normal.  Therefore, these results


      should be interpreted with caution.  A step-wise


      logistic regression was performed looking at


      demographic factors, statin use, and baseline


      disease.  As expected, the most significant factor


      in this analysis was ximelagatran treatment with an


      odds ration of 6.82.


                Other factors that demonstrated


      statistical significance all had an odds ratio of




      less than 2.  These includes patients post-ACS,


      patients being treated for an acute venous


      thromboembolic event, body mass index less than 25


      kilograms per meter squared, and female gender.


      Statins and creatinine clearance were not


      identified as significant factors.


                The variable of ALT greater than 3 times


      the upper limit of normal is generally asymptomatic


      and reversible.  Therefore, this analysis does not


      allow a prediction for those at risk for severe


      liver injury.  We are, therefore, recommending ALT


      testing for everyone who starts long-term treatment


      with ximelagatran.


                Now let's look at the adverse event data


      from these patients.  No difference is noted


      between groups for clinical hepatobiliary adverse




      T2B                      I will now briefly review


      three selected cases in the group of patients with


      concomitant increase in ALT and bilirubin


      associated with ximelagatran.  These cases were


      selected by the FDA as three deaths with associated




      severe liver injury.  The first two cases occurred


      on the first algorithm, and the third case on the


      second more conservative algorithm.  The second and


      third case did not demonstrate compliance with the


      algorithm in effect at the time.  The deaths in all


      three cases are also confounded by other factors.


                In the first two cases, the ALT and


      bilirubin increase was unexplained, and the


      terminal event in both cases was a GI bleeding




                The first patient, an 80-year-old male,


      had a hepatic biopsy with documented hepatic


      necrosis about 1 month before death.  This patient


      had evidence of decreased hepatic function.


      However, the ALT was recovering when he died from a


      perforated duodenal ulcer.  This patient had been


      on prednisone.


                The second case presented hypertensive to


      the hospital with an elevated ALT of 11 times the


      upper limit of normal and a bilirubin of 1.4 times


      the upper limit of normal after missing two weekly


      tests for an elevated ALT.  The INR was 3.4 and the




      APTT was 69 seconds.  His last dose of ximelagatran


      had been earlier that evening.  The patient had a


      prior history of duodenal ulcer and Bilroth II


      anastomosis with bleeding at the site detected on


      this admission.  During the 24 hours from the


      admission to death, the patient received massive


      transfusions.  During this time his bilirubin


      increased from 1.4 times the upper limit of normal


      to 9.4 times the upper limit of normal, with 50


      percent noted as indirect bilirubin.  At the time


      of death, the bilirubin was 7.3 times the upper


      limit of normal and the ALT less than 2 times the


      upper limit of normal.


                The third case was a death due to


      fulminant reactivation hepatitis B with an elevated


      ALT upon study initiation.  This patient was on two


      immunosuppressive drugs:  prednisone and


      azathioprine.  Ximelagatran was not discontinued


      when the ALT reached greater than 5 times the upper


      limit of normal as recommended.  The patient had a


      rapid and fulminant course attributed to the


      hepatitis B.  However, the investigator could not




      rule out that the drug did not contribute to the


      fulminant course.


                To summarize the hepatic findings, ALT


      elevations greater than 3 times the upper limit of


      normal occurred in 7.9 percent of


      ximelagatran-treated patients, occurring primarily


      within the first 6 months.  The elevations were


      typically asymptomatic and reversible, without any


      evidence of an immunoallergic reaction.  An


      incidence of 0.5 percent of concurrent ALT greater


      than 3 times the upper limit of normal and


      bilirubin greater than 2 times the upper limit of


      normal was observed.  Exposure response suggests


      that exposure is not predictive of individual risk


      of transaminase elevation, and no patient subset


      was identified to be at higher risk of developing


      severe hepatic injury.


                Based on the data, we are proposing ALT


      testing in the label reflecting the more


      conservative testing schedule used in clinical


      trials.  To make sure that ALT testing becomes the


      standard of care with ximelagatran, we also




      submitted a risk minimization plan which set out


      our initial proposals to support ALT testing in


      practice.  This proposal was developed after


      extensive external consultation and field testing,


      but we recognize that it may need to be developed


      further in the best interests of ensuring patient


      safety.  We have a meeting arranged with FDA on


      this topic in the near future.


                A few comments on the principles of our


      Risk Minimization Action Plan.  The ultimate goal


      of the plan is to prevent any hepatic failure


      caused by treatment with ximelagatran.  To do this,


      the Risk MAP will help to ensure compliance with


      labeled ALT testing recommendations.  This proposal


      was developed to provide access to ximelagatran by


      those patients who will benefit while minimizing


      risk.  It targets patients, physicians, and


      pharmacists.  It has a strong educational focus and


      is enhanced with practice management tools and


      special packaging.  In addition, following


      discussions with FDA, AstraZeneca will be proposing


      additional enhancements to ensure our ALT testing




      recommendations are followed.  Finally, we have


      proposed continuous evaluation of program




                AstraZeneca understands that the full


      benefit of ximelagatran can only be realized if it


      is used in accordance with the labeled recommendations, and


      to that end we are committed to


      developing the specifics of the program in


      consultation with the agency.


                To complete the assessment of safety, we


      will finish with the overall mortality in the


      long-term exposure pool to get an overview of risk.


      The patient population was primarily an elderly


      population with multiple comorbidities and


      concurrent medications.  Despite an increase in ALT


      in the ximelagatran-treated patients, no difference


      in all-cause mortality was noted.  Mortality was


      similar in the ximelagatran group compared to


      patients on placebo, patients on placebo plus


      aspirin, and patients on warfarin.


                Let me finish by summarizing the


      benefit/risk comments for each indication. 




      Ximelagatran prevented venous thromboembolism


      and/or all-cause mortality compared with warfarin


      in total knee replacement surgery with a number


      needed to treat of 12.  No difference was seen in


      bleeding, transfusions, or surgical outcome.


                Ximelagatran demonstrated clear benefit


      over placebo with a number needed to treat of 10 in


      the long-term prevention of recurrent VTE events.


      This included a clinically important reduction in


      pulmonary embolus, a condition that can result in


      serious morbidity and mortality.  The incidence of


      bleeding was comparable to placebo.


                Ximelagatran was as effective as warfarin


      in reducing the risk of stroke and other


      thromboembolic events in patient with atrial


      fibrillation.  Bleeding was lower on ximelagatran.


      With regards to the hepatic findings, while the


      risk per year for stroke or venous thromboembolism


      is continuous, the risk for an ALT rise and


      subsequent severe liver injury is limited primarily


      to the first 6 months of ximelagatran therapy.  But


      the protection from a thrombotic event by




      ximelagatran is continuous and consistent over




                To aid effective management of the hepatic


      risk, ALT testing will be recommended in our


      proposed labeling, and in addition, we have


      submitted a Risk Minimization Action Plan which we


      will discuss further with the FDA.


                We conclude that ximelagatran, the first


      new oral anticoagulant in over 50 years, does have


      a positive benefit/risk in each proposed indication


      provided that the drug is used properly.  We look


      forward to your comments and further dialogue with


      the agency.


                Thank you.  I'll take questions.


                DR. BORER:  Dr. Sheth, I think we need to


      take a break.  I've been chastised when we haven't


      done that.  So we'll take a 10-minute break, and


      then we'll go on to the questions of fact about the


      safety data, and I think we can then go on to


      Jonathan Halperin's presentation, and we'll just


      make up the remaining FDA time after the public


      comments later so that you can get your whole




      presentation in.


                So we'll take a 10-minute break right now.


      Look at your watch because 10 minutes from now


      we're going to start again.




                DR. BORER:  While everybody is getting


      back in here and sitting down--or not sitting


      down--let me raise an issue for you to begin to


      think about as people are coming back in.


                Steve Nissen asked earlier about


      pharmacological evidence of rebound, and there


      didn't appear to be significant rebound, although I


      don't know what that means in the context of


      studies with limited power.  But there didn't seem


      to be significant rebound of pharmacological


      effects, although the follow-up, as I recall, was


      relatively short.  So we don't know about late


      pharmacological changes.  But as I look at these


      data from each of the trials, I'm struck with a


      difference between the on-treatment and


      post-treatment frequency of major adverse


      cardiovascular events that I'd like to hear some




      discussion about from you.  Is this real or is it


      not?  That is that if you look at the number of


      myocardial infarctions or other cardiac events that


      occurred on ximelagatran versus the comparator, the


      numbers were different but not all that different.


      It depended on the trial.  It varied from trial to


      trial, and we can talk about that potential adverse


      event disparity later.  But I'm concerned or I want


      to ask about something else.


                If you look at the number of events that


      occurred on-treatment, the numbers were relatively


      close one way or the other from trial to trial to


      trial.  If you look at the numbers that occurred


      post-treatment, the proportion of patients who had


      events on ximelagatran in the post-treatment period


      was greater as a percentage of the whole than was


      the case for any of the comparators.  The


      post-treatment events on warfarin or on placebo


      were fewer as a proportion of the whole of the


      total number of events in those comparator groups


      than was the case with ximelagatran, and in some


      cases the post-treatment events were more frequent




      than the on-treatment events with ximelagatran.


      That's an observation.


                Have you noted that?  And is that true?


      And do you have anything to say about it?


                DR. SHETH:  The numbers differ a little


      bit between the different patient groups.  So let


      me start first with the long-term exposure pool and


      some of the specific populations within that pool.


                If we can take a look again at the--and


      we're talking coronary events, Dr. Borer?


                DR. BORER:  Yes, we can limit it to


      coronary events, however they're defined.


                DR. SHETH:  What you see is that, you're


      right, there is a difference--I'm sorry.  Let's put


      that up.  You do see an increase--and these are


      both myocardial infarctions and total other


      coronary artery disease events other than MI


      compared in the VTE treatment and the VTE extended


      prophylaxis compared to warfarin.  But these


      numbers are actually quite small.  We're talking


      about a total of 16 patients here, 3, 16, 10, et


      cetera, versus 1, 0, 12, 3.




                If we take a look at two populations where


      you might say that the risk is actually increased,


      the atrial fib group had higher incidence of both


      diabetes, hypertension, for example.  You don't see


      that--those events, again, plus that in the


      post-acute coronary syndrome population, which is


      certainly a high-risk group for events.  Can we


      take a look at the next slide?  I'll come back to


      the after-treatment in a second.


                Ninety percent of the MIs--and this was


      during the trials--occurred in these two settings,


      and you don't see a difference there, and you see a


      benefit on treatment with ximelagatran.


                But if we take a look, let's say, at the


      post-acute coronary syndrome population, again, a


      higher-risk group, after treatment stopped, the


      incidence between those two groups was about 1.5


      percent--I think it's about 1.5 percent in both




                DR. BORER:  Okay.  I'm sure you're right,


      and the data you just showed I think are very


      reassuring, and I think we all saw them in the book




      here.  But, again, I'm making a slightly different


      point, and maybe the data aren't available or


      aren't sufficient to draw a firm conclusion about




                What I am talking about is the proportion


      of coronary events that occurred after stopping


      treatment on ximelagatran as a percentage of the


      total number of events compared with the portion


      that occurred after stopping treatment with


      warfarin or placebo as a percentage of the total


      number of events in those groups.  I believe that


      the proportion of events that occur post-treatment


      is higher in the ximelagatran groups across all the


      trials, if you look at trial after trial, than is


      the case for the comparators, which raises some


      question about the possibility of a rebound


      phenomenon or something else, some other


      pathophysiological process that's being allowed to


      happen or occurring because of the use of the drug


      once it's stopped.


                DR. SHETH:  I understand what you're


      asking.  We don't have that specific analysis, so I




      won't be able to address it at this moment.  You're


      asking for those proportions of patients after they


      stop treatment over the total number of events, and


      right now I don't have that analysis.


                DR. BORER:  Okay.  You can pull it


      together later, but it's in the books.  If you look


      at the data that are presented, if you look at the


      numbers, that sort of jumps out at you.  So you may


      want to look at that, and you can talk about it


      after lunch or something.


                DR. SHETH:  Okay.


                DR. BORER:  Okay.  Well, why don't we go


      on and see--Alan?


                DR. HIRSCH:  Well, just one comment to


      follow up Jeff, and if you're able to provide that


      after lunch, I specifically would ask you provide


      that not in the ACS population, because the


      population that will be exposed to this if this


      drug comes to market that really is vulnerable that


      I'm concerned about is that non-ACS population.


                DR. SHETH:  Okay.


                DR. HIRSCH:  I don't want that to be a




      Band-aid for a potential adverse effect.


                DR. BORER:  Steve and then Bill.


                DR. NISSEN:  Just so you understand what


      we're concerned about--and several of us have made


      this observation--it is that because ximelagatran


      is a short-acting agent compared to Coumadin, our


      worry is that when you stop the drug, there's some


      phenomenon that goes on for a few days or a few


      weeks in which a patient has increased vulnerability and


      that that is the explanation for the


      excess cardiovascular events.  And we want to


      understand whether you have some response to that


      that we can factor into our thinking.


                DR. SHETH:  Can I ask, would it help the


      committee to take a look at other thrombotic events


      in terms of incident or rebound phenomena?  Because


      certainly patients who are usually at risk for


      venous events might typically get those kind of


      events.  Would that help--


                DR. NISSEN:  It only helps a little bit.


      The problem is that the pathophysiology of arterial


      and venous events are different.




                DR. SHETH:  Right.


                DR. NISSEN:  And so, you know, it appears


      that there is this excess of arterial thrombotic


      events post-treatment, and we're trying to


      understand that in order to factor that into the


      thinking here of the committee.


                DR. SHETH:  Right, although in maybe the


      treatment and prevention groups--


                DR. NISSEN:  Yes, yes.  I have another


      question, and forgive me for this, but I have to


      probe on something that I think is important.  If


      you could put up Slide No. CE-19, please?  I see


      these patients that are going to have knee


      replacement all the time in consultation.  They


      almost all get sent for cardiac clearance because


      they're older and they have a lot of cardiovascular


      risk factors, and I'll bet the other cardiologists


      at this table, like Jeff probably sees plenty of


      these as well.  And so when I see them, there are


      three things that I worry about.  I worry about, of


      course, them dying.  I worry about them having a


      pulmonary embolus.  And I worry about them having a




      myocardial infarction.


                And so, you know, to do the simple math


      here, which is what all of us are kind of looking


      at, if you look at the serious endpoints, the


      feared complications, what you see is--in EXULT A


      and B, you see three plus six is nine events with


      ximelagatran, and you see eight events here, PE or


      death, with warfarin.


                If you now put up Slide No.--


                DR. FLEMING:  Steve, just before you go,


      those nine and eight are 15 and 10 in the FDA


      briefing document.  It's worse than this.  It's 15


      and 10.


                DR. NISSEN:  Okay.  I'm trying to be--you


      know, not make this any more painful than it has to




                Now let's look at Slide CS-14, so we'll


      take nine and eight, so CS-14, and now I look at


      myocardial infarction, and it's 16 to 4.  And so


      when you put it together, you know, you see that


      the really serious events, the bad things that can


      happen to that patient I'm seeing in consultation




      look a lot worse on ximelagatran than warfarin.


      And so one has to ask the question:  Does it really


      look as good as it looks?


                And so what are your thoughts about this?


      I mean, MI is as bad an outcome as PE, isn't it?


                DR. SHETH:  Yes, it is.  In considering


      those numbers, I won't dispute how--the numbers


      that we just looked at, they are higher in the


      ximelagatran group compared to the warfarin group


      in the orthopedic surgery population.  The only


      comment I'd make is that, unfortunately--those are


      really small numbers, and the question is:  Is this


      a really--a true difference?  And I would


      anticipate that if it was a true effect that we


      would really see a significant effect in the


      long-term group just because it's so much larger.


                We also have another study that we started


      to do in extended prophylaxis in orthopedic


      surgery, so we'll be able to collect more data in


      that study as well.  But, again, the numbers are


      small so it's hard to know if this is a true


      difference or not.




                DR. NISSEN:  But, of course, the


      difference in the long-term studies is that this is


      one where you get the short-term administration,


      then you withdraw the drug, and so it speaks more


      to this question of an acute rebound sort of


      phenomenon.  I mean, I hope you can appreciate why


      it's something that really struck many of us on the


      committee as being a problem.


                DR. BORER:  It's also a potentially


      remediable problem, so it's important that you


      should know about it.


                Bill, and then Beverly.


                DR. HIATT:  Yes, just to follow up on


      that, it does seem like the surgical population may


      not be the same as the long-term treatment


      population, and the concept of risk


      occurring--excess risk occurring in that population


      is very real.


                Then the other question I have is, turning


      to the SPORTIF IV data, you didn't present that in


      any of your safety data.  Is that correct?


                DR. SHETH:  The SPORTIF II and IV data are




      actually pooled in the atrial fibrillation pool


      that we performed.  So it included that Phase II


      trial, yes.


                DR. HIATT:  If you look at page 97 of the


      briefing document, there are several phases to


      SPORTIF IV, and I count a total of 17 deaths on


      treatment versus warfarin is four.  So you're


      saying those deaths are included in the overall


      safety data you presented?


                DR. SHETH:  They are included, but I'll


      just point out that there are about 2 to 3 times


      more patients on--3 times more patients on


      ximelagatran than on warfarin in SPORTIF IV.  So


      they're not balanced groups.  The denominators are


      not balanced.


                DR. HIATT:  Correct.


                DR. SHETH:  But those deaths are included


      in the atrial fib pool and consequently in the


      long-term exposure pool.


                DR. HIATT:  Okay.


                DR. BORER:  Beverly, and then Dr. Sjogren.


                DR. LORELL:  To follow up on this concept




      of potential rebound--


                DR. SHETH:  Can you speak louder?


                DR. LORELL:  Yes, I can.  To follow up on


      the issues that were raised about potential rebound


      in the post-surgical population, can you enlighten


      us as to how investigators were instructed to use


      or no instructions on aspirin?  Was aspirin


      deliberately not used in that surgical population?


      And then were there any instructions at the


      termination of treatment?


                DR. SHETH:  Let me ask Dr. Berkowitz, who


      was the physician for those studies, to describe


      the use of aspirin instructions for the surgical




                DR. BERKOWITZ:  Scott Berkowitz,


      AstraZeneca.  I didn't get the second part.  The


      first part was that aspirin was precluded, kept to


      a minimum, and patients weren't to be on it




                DR. LORELL:  So I think the second part,


      was there a strategy in that trial when the study


      drug was stopped about reinstatement of aspirin in




      patients who had risk factors?  You know, the point


      that Dr. Borer made, this is a group with rich risk




                DR. BERKOWITZ:  I'm sorry.  I think I got


      all your question.  The studies were designed to


      leave to the discretion of the investigators to put


      the patients back on the medicine, so we did not


      prespecify how to do that.


                DR. LORELL:  Okay.  And related to that,


      have you done any studies after withdrawal of the


      drug to look at what happens to platelet function?


                DR. BERKOWITZ:  In our clinical trials for


      VTE and orthopedic surgery and I believe in the


      atrial fibrillation trials, we did not do any


      platelet studies.


                DR. SHETH:  I can mention we actually


      looked and did an analysis of patients both on


      ximelagatran, on aspirin and off aspirin, for


      events in the atrial fibrillation pool.  If you're


      interested, we can show that if that would be


      helpful.  And this is not exactly the same as the


      patients who discontinued after orthopedic surgery.




      But if you're concerned about any increased


      beneficial effect--let's see.  Actually, what you


      see is that there is an incremental benefit in


      patients who are on aspirin, but you see that same


      benefit on warfarin, and you don't see a difference


      of the effect between the two anticoagulants when


      aspirin is added.


                DR. BORER:  The question that Beverly is


      asking, though, is what about after you've stopped


      the ximelagatran and the warfarin.  In that period,


      were people still on aspirin or were they not?  And


      did the fact that they were or weren't have any


      impact on the post-treatment events?


                DR. SHETH:  We didn't make specific


      recommendations after the trial.  They were to go


      on their regular medications per their physician.


                DR. BORER:  Okay.  Dr. Sjogren, then Alan,


      then Ron.


                DR. SJOGREN:  My question pertains to the


      potential hepatic toxicity, and I have a couple of


      questions.  One is you are proposing to follow up


      patients with ALTs, and then if they go over 2




      times the upper limit of normal, to follow up a


      little more closely and eventually discontinue the


      drug.  I'd like to know what kind of information do


      you have in the patients that you followed up that


      developed the ALT abnormality to back up that kind


      of recommendation.  That's one question.


                The second question is:  Do you have any


      information on patients with chronic liver disease


      that are treated with this medication?  What


      happens to them?


                And one request.  Do you have slides of


      the liver biopsy that was done that we can look at?


                DR. BORER:  Before you begin to answer,


      let me just state a rule here.  We're not asking


      you to tell us your algorithm for following


      patients.  Dr. Sjogren is just asking about the


      data that might be used to inform the development


      of such an algorithm and then the issue of the


      chronic use and the slides.


                DR. SHETH:  Okay.  Let me answer the


      latter two questions first, and I'm going to have


      to ask for a clarification on that first one.




                We do not have a slide of the hepatic


      biopsy right now.  The chronic disease, because we


      identified early in the Phase II trial, SPORTIF II,


      that there was this asymptomatic transaminase


      increase, we actually excluded patients who had


      known hepatic disease from the trials, as well as


      patients who had an ALT above 2 times the upper


      limit of normal.  So that to the best of our


      knowledge, patients who--with the exception of the


      reactivation hepatitis B, should not have, in fact,


      been enrolled in the trial, and we would, in fact,


      propose a contraindication for those patients.  So


      we don't have data to understand the safety in that




                In terms of follow-up, are you asking


      me--you want to know what did we do to follow up


      all patients who had an elevation or what their


      outcome was?